Rynson W. H. Lau

Tianyu Huang, Bowen Dong, Jiaying Lin et al.

Mirror detection aims to identify the mirror regions in the given input image. Existing works mainly focus on integrating the semantic features and structural features to mine specific relations between mirror and non-mirror regions, or introducing mirror properties like depth or chirality to help analyze the existence of mirrors. In this work, we observe that a real object typically forms a loose symmetry relationship with its corresponding reflection in the mirror, which is beneficial in distinguishing mirrors from real objects. Based on this observation, we propose a dual-path Symmetry-Aware Transformer-based mirror detection Network (SATNet), which includes two novel modules: Symmetry-Aware Attention Module (SAAM) and Contrast and Fusion Decoder Module (CFDM). Specifically, we first adopt a transformer backbone to model global information aggregation in images, extracting multi-scale features in two paths. We then feed the high-level dual-path features to SAAMs to capture the symmetry relations. Finally, we fuse the dual-path features and refine our prediction maps progressively with CFDMs to obtain the final mirror mask. Experimental results show that SATNet outperforms both RGB and RGB-D mirror detection methods on all available mirror detection datasets. Codes and trained models are available at: https://github.com/tyhuang0428/SATNet.

Zhanghan Ke, Chunyi Sun, Lei Zhu et al.

Recent works on image harmonization solve the problem as a pixel-wise image translation task via large autoencoders. They have unsatisfactory performances and slow inference speeds when dealing with high-resolution images. In this work, we observe that adjusting the input arguments of basic image filters, e.g., brightness and contrast, is sufficient for humans to produce realistic images from the composite ones. Hence, we frame image harmonization as an image-level regression problem to learn the arguments of the filters that humans use for the task. We present a Harmonizer framework for image harmonization. Unlike prior methods that are based on black-box autoencoders, Harmonizer contains a neural network for filter argument prediction and several white-box filters (based on the predicted arguments) for image harmonization. We also introduce a cascade regressor and a dynamic loss strategy for Harmonizer to learn filter arguments more stably and precisely. Since our network only outputs image-level arguments and the filters we used are efficient, Harmonizer is much lighter and faster than existing methods. Comprehensive experiments demonstrate that Harmonizer surpasses existing methods notably, especially with high-resolution inputs. Finally, we apply Harmonizer to video harmonization, which achieves consistent results across frames and 56 fps at 1080P resolution. Code and models are available at: https://github.com/ZHKKKe/Harmonizer.

Xin Tian, Ke Xu, Xin Yang et al.

The saliency ranking task is recently proposed to study the visual behavior that humans would typically shift their attention over different objects of a scene based on their degrees of saliency. Existing approaches focus on learning either object-object or object-scene relations. Such a strategy follows the idea of object-based attention in Psychology, but it tends to favor those objects with strong semantics (e.g., humans), resulting in unrealistic saliency ranking. We observe that spatial attention works concurrently with object-based attention in the human visual recognition system. During the recognition process, the human spatial attention mechanism would move, engage, and disengage from region to region (i.e., context to context). This inspires us to model the region-level interactions, in addition to the object-level reasoning, for saliency ranking. To this end, we propose a novel bi-directional method to unify spatial attention and object-based attention for saliency ranking. Our model includes two novel modules: (1) a selective object saliency (SOS) module that models objectbased attention via inferring the semantic representation of the salient object, and (2) an object-context-object relation (OCOR) module that allocates saliency ranks to objects by jointly modeling the object-context and context-object interactions of the salient objects. Extensive experiments show that our approach outperforms existing state-of-theart methods. Our code and pretrained model are available at https://github.com/GrassBro/OCOR.

Tianyu Huang, Bowen Dong, Yunhan Yang et al.

Pre-training across 3D vision and language remains under development because of limited training data. Recent works attempt to transfer vision-language pre-training models to 3D vision. PointCLIP converts point cloud data to multi-view depth maps, adopting CLIP for shape classification. However, its performance is restricted by the domain gap between rendered depth maps and images, as well as the diversity of depth distributions. To address this issue, we propose CLIP2Point, an image-depth pre-training method by contrastive learning to transfer CLIP to the 3D domain, and adapt it to point cloud classification. We introduce a new depth rendering setting that forms a better visual effect, and then render 52,460 pairs of images and depth maps from ShapeNet for pre-training. The pre-training scheme of CLIP2Point combines cross-modality learning to enforce the depth features for capturing expressive visual and textual features and intra-modality learning to enhance the invariance of depth aggregation. Additionally, we propose a novel Dual-Path Adapter (DPA) module, i.e., a dual-path structure with simplified adapters for few-shot learning. The dual-path structure allows the joint use of CLIP and CLIP2Point, and the simplified adapter can well fit few-shot tasks without post-search. Experimental results show that CLIP2Point is effective in transferring CLIP knowledge to 3D vision. Our CLIP2Point outperforms PointCLIP and other self-supervised 3D networks, achieving state-of-the-art results on zero-shot and few-shot classification.

Ruozhen He, Qihua Dong, Jiaying Lin et al.

Existing camouflaged object detection (COD) methods rely heavily on large-scale datasets with pixel-wise annotations. However, due to the ambiguous boundary, annotating camouflage objects pixel-wisely is very time-consuming and labor-intensive, taking ~60mins to label one image. In this paper, we propose the first weakly-supervised COD method, using scribble annotations as supervision. To achieve this, we first relabel 4,040 images in existing camouflaged object datasets with scribbles, which takes ~10s to label one image. As scribble annotations only describe the primary structure of objects without details, for the network to learn to localize the boundaries of camouflaged objects, we propose a novel consistency loss composed of two parts: a cross-view loss to attain reliable consistency over different images, and an inside-view loss to maintain consistency inside a single prediction map. Besides, we observe that humans use semantic information to segment regions near the boundaries of camouflaged objects. Hence, we further propose a feature-guided loss, which includes visual features directly extracted from images and semantically significant features captured by the model. Finally, we propose a novel network for COD via scribble learning on structural information and semantic relations. Our network has two novel modules: the local-context contrasted (LCC) module, which mimics visual inhibition to enhance image contrast/sharpness and expand the scribbles into potential camouflaged regions, and the logical semantic relation (LSR) module, which analyzes the semantic relation to determine the regions representing the camouflaged object. Experimental results show that our model outperforms relevant SOTA methods on three COD benchmarks with an average improvement of 11.0% on MAE, 3.2% on S-measure, 2.5% on E-measure, and 4.4% on weighted F-measure.

Zhanghan Ke, Yuhao Liu, Lei Zhu et al.

In this paper, we present a Neural Preset technique to address the limitations of existing color style transfer methods, including visual artifacts, vast memory requirement, and slow style switching speed. Our method is based on two core designs. First, we propose Deterministic Neural Color Mapping (DNCM) to consistently operate on each pixel via an image-adaptive color mapping matrix, avoiding artifacts and supporting high-resolution inputs with a small memory footprint. Second, we develop a two-stage pipeline by dividing the task into color normalization and stylization, which allows efficient style switching by extracting color styles as presets and reusing them on normalized input images. Due to the unavailability of pairwise datasets, we describe how to train Neural Preset via a self-supervised strategy. Various advantages of Neural Preset over existing methods are demonstrated through comprehensive evaluations. Notably, Neural Preset enables stable 4K color style transfer in real-time without artifacts. Besides, we show that our trained model can naturally support multiple applications without fine-tuning, including low-light image enhancement, underwater image correction, image dehazing, and image harmonization. Project page with demos: https://zhkkke.github.io/NeuralPreset .

Zhenwei Wang, Nanxuan Zhao, Gerhard Hancke et al.

Adjusting the photo color to associate with some design elements is an essential way for a graphic design to effectively deliver its message and make it aesthetically pleasing. However, existing tools and previous works face a dilemma between the ease of use and level of expressiveness. To this end, we introduce an interactive language-based approach for photo recoloring, which provides an intuitive system that can assist both experts and novices on graphic design. Given a graphic design containing a photo that needs to be recolored, our model can predict the source colors and the target regions, and then recolor the target regions with the source colors based on the given language-based instruction. The multi-granularity of the instruction allows diverse user intentions. The proposed novel task faces several unique challenges, including: 1) color accuracy for recoloring with exactly the same color from the target design element as specified by the user; 2) multi-granularity instructions for parsing instructions correctly to generate a specific result or multiple plausible ones; and 3) locality for recoloring in semantically meaningful local regions to preserve original image semantics. To address these challenges, we propose a model called LangRecol with two main components: the language-based source color prediction module and the semantic-palette-based photo recoloring module. We also introduce an approach for generating a synthetic graphic design dataset with instructions to enable model training. We evaluate our model via extensive experiments and user studies. We also discuss several practical applications, showing the effectiveness and practicality of our approach. Code and data for this paper are at: https://zhenwwang.github.io/langrecol.

Ruozhen He, Jiaying Lin, Rynson W. H. Lau

We present HetNet (Multi-level \textbf{Het}erogeneous \textbf{Net}work), a highly efficient mirror detection network. Current mirror detection methods focus more on performance than efficiency, limiting the real-time applications (such as drones). Their lack of efficiency is aroused by the common design of adopting homogeneous modules at different levels, which ignores the difference between different levels of features. In contrast, HetNet detects potential mirror regions initially through low-level understandings (\textit{e.g.}, intensity contrasts) and then combines with high-level understandings (contextual discontinuity for instance) to finalize the predictions. To perform accurate yet efficient mirror detection, HetNet follows an effective architecture that obtains specific information at different stages to detect mirrors. We further propose a multi-orientation intensity-based contrasted module (MIC) and a reflection semantic logical module (RSL), equipped on HetNet, to predict potential mirror regions by low-level understandings and analyze semantic logic in scenarios by high-level understandings, respectively. Compared to the state-of-the-art method, HetNet runs 664$\%$ faster and draws an average performance gain of 8.9$\%$ on MAE, 3.1$\%$ on IoU, and 2.0$\%$ on F-measure on two mirror detection benchmarks.

Haiyang Mei, Xin Yang, Letian Yu et al.

Glass is very common in our daily life. Existing computer vision systems neglect it and thus may have severe consequences, e.g., a robot may crash into a glass wall. However, sensing the presence of glass is not straightforward. The key challenge is that arbitrary objects/scenes can appear behind the glass. In this paper, we propose an important problem of detecting glass surfaces from a single RGB image. To address this problem, we construct the first large-scale glass detection dataset (GDD) and propose a novel glass detection network, called GDNet-B, which explores abundant contextual cues in a large field-of-view via a novel large-field contextual feature integration (LCFI) module and integrates both high-level and low-level boundary features with a boundary feature enhancement (BFE) module. Extensive experiments demonstrate that our GDNet-B achieves satisfying glass detection results on the images within and beyond the GDD testing set. We further validate the effectiveness and generalization capability of our proposed GDNet-B by applying it to other vision tasks, including mirror segmentation and salient object detection. Finally, we show the potential applications of glass detection and discuss possible future research directions.

Tao Yan, Mingyue Li, Bin Li et al.

Existing deraining methods focus mainly on a single input image. However, with just a single input image, it is extremely difficult to accurately detect and remove rain streaks, in order to restore a rain-free image. In contrast, a light field image (LFI) embeds abundant 3D structure and texture information of the target scene by recording the direction and position of each incident ray via a plenoptic camera. LFIs are becoming popular in the computer vision and graphics communities. However, making full use of the abundant information available from LFIs, such as 2D array of sub-views and the disparity map of each sub-view, for effective rain removal is still a challenging problem. In this paper, we propose a novel method, 4D-MGP-SRRNet, for rain streak removal from LFIs. Our method takes as input all sub-views of a rainy LFI. To make full use of the LFI, it adopts 4D convolutional layers to simultaneously process all sub-views of the LFI. In the pipeline, the rain detection network, MGPDNet, with a novel Multi-scale Self-guided Gaussian Process (MSGP) module is proposed to detect high-resolution rain streaks from all sub-views of the input LFI at multi-scales. Semi-supervised learning is introduced for MSGP to accurately detect rain streaks by training on both virtual-world rainy LFIs and real-world rainy LFIs at multi-scales via computing pseudo ground truths for real-world rain streaks. We then feed all sub-views subtracting the predicted rain streaks into a 4D convolution-based Depth Estimation Residual Network (DERNet) to estimate the depth maps, which are later converted into fog maps. Finally, all sub-views concatenated with the corresponding rain streaks and fog maps are fed into a powerful rainy LFI restoring model based on the adversarial recurrent neural network to progressively eliminate rain streaks and recover the rain-free LFI.

Yuhao Liu, Qing Guo, Lan Fu et al.

Existing deep learning-based shadow removal methods still produce images with shadow remnants. These shadow remnants typically exist in homogeneous regions with low-intensity values, making them untraceable in the existing image-to-image mapping paradigm. We observe that shadows mainly degrade images at the image-structure level (in which humans perceive object shapes and continuous colors). Hence, in this paper, we propose to remove shadows at the image structure level. Based on this idea, we propose a novel structure-informed shadow removal network (StructNet) to leverage the image-structure information to address the shadow remnant problem. Specifically, StructNet first reconstructs the structure information of the input image without shadows and then uses the restored shadow-free structure prior to guiding the image-level shadow removal. StructNet contains two main novel modules: (1) a mask-guided shadow-free extraction (MSFE) module to extract image structural features in a non-shadow-to-shadow directional manner, and (2) a multi-scale feature & residual aggregation (MFRA) module to leverage the shadow-free structure information to regularize feature consistency. In addition, we also propose to extend StructNet to exploit multi-level structure information (MStructNet), to further boost the shadow removal performance with minimum computational overheads. Extensive experiments on three shadow removal benchmarks demonstrate that our method outperforms existing shadow removal methods, and our StructNet can be integrated with existing methods to improve them further.

Tianyu Huang, Yihan Zeng, Bowen Dong et al.

Recent works learn 3D representation explicitly under text-3D guidance. However, limited text-3D data restricts the vocabulary scale and text control of generations. Generators may easily fall into a stereotype concept for certain text prompts, thus losing open-vocabulary generation ability. To tackle this issue, we introduce a conditional 3D generative model, namely TextField3D. Specifically, rather than using the text prompts as input directly, we suggest to inject dynamic noise into the latent space of given text prompts, i.e., Noisy Text Fields (NTFs). In this way, limited 3D data can be mapped to the appropriate range of textual latent space that is expanded by NTFs. To this end, an NTFGen module is proposed to model general text latent code in noisy fields. Meanwhile, an NTFBind module is proposed to align view-invariant image latent code to noisy fields, further supporting image-conditional 3D generation. To guide the conditional generation in both geometry and texture, multi-modal discrimination is constructed with a text-3D discriminator and a text-2.5D discriminator. Compared to previous methods, TextField3D includes three merits: 1) large vocabulary, 2) text consistency, and 3) low latency. Extensive experiments demonstrate that our method achieves a potential open-vocabulary 3D generation capability.

Zhenwei Wang, Tengfei Wang, Zexin He et al.

In 3D modeling, designers often use an existing 3D model as a reference to create new ones. This practice has inspired the development of Phidias, a novel generative model that uses diffusion for reference-augmented 3D generation. Given an image, our method leverages a retrieved or user-provided 3D reference model to guide the generation process, thereby enhancing the generation quality, generalization ability, and controllability. Our model integrates three key components: 1) meta-ControlNet that dynamically modulates the conditioning strength, 2) dynamic reference routing that mitigates misalignment between the input image and 3D reference, and 3) self-reference augmentations that enable self-supervised training with a progressive curriculum. Collectively, these designs result in a clear improvement over existing methods. Phidias establishes a unified framework for 3D generation using text, image, and 3D conditions with versatile applications.

Youjun Zhao, Jiaying Lin, Shuquan Ye et al.

Open-vocabulary 3D scene understanding (OV-3D) aims to localize and classify novel objects beyond the closed set of object classes. However, existing approaches and benchmarks primarily focus on the open vocabulary problem within the context of object classes, which is insufficient in providing a holistic evaluation to what extent a model understands the 3D scene. In this paper, we introduce a more challenging task called Generalized Open-Vocabulary 3D Scene Understanding (GOV-3D) to explore the open vocabulary problem beyond object classes. It encompasses an open and diverse set of generalized knowledge, expressed as linguistic queries of fine-grained and object-specific attributes. To this end, we contribute a new benchmark named \textit{OpenScan}, which consists of 3D object attributes across eight representative linguistic aspects, including affordance, property, and material. We further evaluate state-of-the-art OV-3D methods on our OpenScan benchmark and discover that these methods struggle to comprehend the abstract vocabularies of the GOV-3D task, a challenge that cannot be addressed simply by scaling up object classes during training. We highlight the limitations of existing methodologies and explore promising directions to overcome the identified shortcomings.

Tianyu Huang, Yihan Zeng, Zhilu Zhang et al.

3D generation has raised great attention in recent years. With the success of text-to-image diffusion models, the 2D-lifting technique becomes a promising route to controllable 3D generation. However, these methods tend to present inconsistent geometry, which is also known as the Janus problem. We observe that the problem is caused mainly by two aspects, i.e., viewpoint bias in 2D diffusion models and overfitting of the optimization objective. To address it, we propose a two-stage 2D-lifting framework, namely DreamControl, which optimizes coarse NeRF scenes as 3D self-prior and then generates fine-grained objects with control-based score distillation. Specifically, adaptive viewpoint sampling and boundary integrity metric are proposed to ensure the consistency of generated priors. The priors are then regarded as input conditions to maintain reasonable geometries, in which conditional LoRA and weighted score are further proposed to optimize detailed textures. DreamControl can generate high-quality 3D content in terms of both geometry consistency and texture fidelity. Moreover, our control-based optimization guidance is applicable to more downstream tasks, including user-guided generation and 3D animation. The project page is available at https://github.com/tyhuang0428/DreamControl.

Jiaying Lin, Yuen-Hei Yeung, Shuquan Ye et al.

Glass surfaces are becoming increasingly ubiquitous as modern buildings tend to use a lot of glass panels. This, however, poses substantial challenges to the operations of autonomous systems such as robots, self-driving cars, and drones, as these glass panels can become transparent obstacles to navigation. Existing works attempt to exploit various cues, including glass boundary context or reflections, as priors. However, they are all based on input RGB images. We observe that the transmission of 3D depth sensor light through glass surfaces often produces blank regions in the depth maps, which can offer additional insights to complement the RGB image features for glass surface detection. In this work, we first propose a large-scale RGB-D glass surface detection dataset, \textit{RGB-D GSD}, for rigorous experiments and future research. It contains 3,009 images, paired with precise annotations, offering a wide range of real-world RGB-D glass surface categories. We then propose a novel glass surface detection framework combining RGB and depth information, with two novel modules: a cross-modal context mining (CCM) module to adaptively learn individual and mutual context features from RGB and depth information, and a depth-missing aware attention (DAA) module to explicitly exploit spatial locations where missing depths occur to help detect the presence of glass surfaces. Experimental results show that our proposed model outperforms state-of-the-art methods.

Rui Song, Jiaying Lin, Rynson W. H. Lau

Video mirror detection has received significant research attention, yet existing methods suffer from limited performance and robustness. These approaches often over-rely on single, unreliable dynamic features, and are typically built on CNNs with limited receptive fields or Transformers with quadratic computational complexity. To address these limitations, we propose a new effective and scalable video mirror detection method, called MirrorMamba. Our approach leverages multiple cues to adapt to diverse conditions, incorporating perceived depth, correspondence and optical. We also introduce an innovative Mamba-based Multidirection Correspondence Extractor, which benefits from the global receptive field and linear complexity of the emerging Mamba spatial state model to effectively capture correspondence properties. Additionally, we design a Mamba-based layer-wise boundary enforcement decoder to resolve the unclear boundary caused by the blurred depth map. Notably, this work marks the first successful application of the Mamba-based architecture in the field of mirror detection. Extensive experiments demonstrate that our method outperforms existing state-of-the-art approaches for video mirror detection on the benchmark datasets. Furthermore, on the most challenging and representative image-based mirror detection dataset, our approach achieves state-of-the-art performance, proving its robustness and generalizability.

Zefan Qu, Ke Xu, Gerhard Petrus Hancke et al.

Neural Radiance Fields (NeRFs) have shown remarkable performances in producing novel-view images from high-quality scene images. However, hand-held low-light photography challenges NeRFs as the captured images may simultaneously suffer from low visibility, noise, and camera shakes. While existing NeRF methods may handle either low light or motion, directly combining them or incorporating additional image-based enhancement methods does not work as these degradation factors are highly coupled. We observe that noise in low-light images is always sharp regardless of camera shakes, which implies an implicit order of these degradation factors within the image formation process. To this end, we propose in this paper a novel model, named LuSh-NeRF, which can reconstruct a clean and sharp NeRF from a group of hand-held low-light images. The key idea of LuSh-NeRF is to sequentially model noise and blur in the images via multi-view feature consistency and frequency information of NeRF, respectively. Specifically, LuSh-NeRF includes a novel Scene-Noise Decomposition (SND) module for decoupling the noise from the scene representation and a novel Camera Trajectory Prediction (CTP) module for the estimation of camera motions based on low-frequency scene information. To facilitate training and evaluations, we construct a new dataset containing both synthetic and real images. Experiments show that LuSh-NeRF outperforms existing approaches. Our code and dataset can be found here: https://github.com/quzefan/LuSh-NeRF.

Lei Zhu, Xinjiang Wang, Wayne Zhang et al.

Convolutions (Convs) and multi-head self-attentions (MHSAs) are typically considered alternatives to each other for building vision backbones. Although some works try to integrate both, they apply the two operators simultaneously at the finest pixel granularity. With Convs responsible for per-pixel feature extraction already, the question is whether we still need to include the heavy MHSAs at such a fine-grained level. In fact, this is the root cause of the scalability issue w.r.t. the input resolution for vision transformers. To address this important problem, we propose in this work to use MSHAs and Convs in parallel \textbf{at different granularity levels} instead. Specifically, in each layer, we use two different ways to represent an image: a fine-grained regular grid and a coarse-grained set of semantic slots. We apply different operations to these two representations: Convs to the grid for local features, and MHSAs to the slots for global features. A pair of fully differentiable soft clustering and dispatching modules is introduced to bridge the grid and set representations, thus enabling local-global fusion. Through extensive experiments on various vision tasks, we empirically verify the potential of the proposed integration scheme, named \textit{GLMix}: by offloading the burden of fine-grained features to light-weight Convs, it is sufficient to use MHSAs in a few (e.g., 64) semantic slots to match the performance of recent state-of-the-art backbones, while being more efficient. Our visualization results also demonstrate that the soft clustering module produces a meaningful semantic grouping effect with only IN1k classification supervision, which may induce better interpretability and inspire new weakly-supervised semantic segmentation approaches. Code will be available at \url{https://github.com/rayleizhu/GLMix}.

Huankang Guan, Ke Xu, Rynson W. H. Lau

Shadow detection is a challenging task as it requires a comprehensive understanding of shadow characteristics and global/local illumination conditions. We observe from our experiment that state-of-the-art deep methods tend to have higher error rates in differentiating shadow pixels from non-shadow pixels in dark regions (ie, regions with low-intensity values). Our key insight to this problem is that existing methods typically learn discriminative shadow features from the whole image globally, covering the full range of intensity values, and may not learn the subtle differences between shadow and non-shadow pixels in dark regions. Hence, if we can design a model to focus on a narrower range of low-intensity regions, it may be able to learn better discriminative features for shadow detection. Inspired by this insight, we propose a novel shadow detection approach that first learns global contextual cues over the entire image and then zooms into the dark regions to learn local shadow representations. To this end, we formulate an effective dark-region recommendation (DRR) module to recommend regions of low-intensity values, and a novel dark-aware shadow analysis (DASA) module to learn dark-aware shadow features from the recommended dark regions. Extensive experiments show that the proposed method outperforms the state-of-the-art methods on three popular shadow detection datasets. Code is available at https://github.com/guanhuankang/ShadowDetection2021.git.

Zaiquan Yang, Yuhao Liu, Gerhard Hancke et al.

Spatio-temporal video grounding (STVG) aims at localizing the spatio-temporal tube of a video, as specified by the input text query. In this paper, we utilize multimodal large language models (MLLMs) to explore a zero-shot solution in STVG. We reveal two key insights about MLLMs: (1) MLLMs tend to dynamically assign special tokens, referred to as \textit{grounding tokens}, for grounding the text query; and (2) MLLMs often suffer from suboptimal grounding due to the inability to fully integrate the cues in the text query (\textit{e.g.}, attributes, actions) for inference. Based on these insights, we propose a MLLM-based zero-shot framework for STVG, which includes novel decomposed spatio-temporal highlighting (DSTH) and temporal-augmented assembling (TAS) strategies to unleash the reasoning ability of MLLMs. The DSTH strategy first decouples the original query into attribute and action sub-queries for inquiring the existence of the target both spatially and temporally. It then uses a novel logit-guided re-attention (LRA) module to learn latent variables as spatial and temporal prompts, by regularizing token predictions for each sub-query. These prompts highlight attribute and action cues, respectively, directing the model's attention to reliable spatial and temporal related visual regions. In addition, as the spatial grounding by the attribute sub-query should be temporally consistent, we introduce the TAS strategy to assemble the predictions using the original video frames and the temporal-augmented frames as inputs to help improve temporal consistency. We evaluate our method on various MLLMs, and show that it outperforms SOTA methods on three common STVG benchmarks. The code will be available at https://github.com/zaiquanyang/LLaVA_Next_STVG.

Dong Liang, Yuhao Liu, Jinyuan Jia et al.

Being able to edit panoramic images is crucial for creating realistic 360° visual experiences. However, existing perspective-based image editing methods fail to model the spatial structure of panoramas. Conventional cube-map decompositions attempt to overcome this problem but inevitably break global consistency due to their mismatch with spherical geometry. Motivated by this insight, we reformulate panoramic editing directly in the equirectangular projection (ERP) domain and present World-Shaper, a unified geometry-aware framework that bridges panoramic generation and editing within a single editing-centric design. To overcome the scarcity of paired data, we adopt a generate-then-edit paradigm, where controllable panoramic generation serves as an auxiliary stage to synthesize diverse paired examples for supervised editing learning. To address geometric distortion, we introduce a geometry-aware learning strategy that explicitly enforces position-aware shape supervision and implicitly internalizes panoramic priors through progressive training. Extensive experiments on our new benchmark, PEBench, demonstrate that our method achieves superior geometric consistency, editing fidelity, and text controllability compared to SOTA methods, enabling coherent and flexible 360° visual world creation with unified editing control. Code, model, and data will be released at our project page: https://world-shaper-project.github.io/

Zhicun Yin, Junjie Chen, Ming Liu et al.

Blind facial image restoration is highly challenging due to unknown complex degradations and the sensitivity of humans to faces. Although existing methods introduce auxiliary information from generative priors or high-quality reference images, they still struggle with identity preservation problems, mainly due to improper feature introduction on detailed textures. In this paper, we focus on effectively incorporating appropriate features from high-quality reference images, presenting a novel blind facial image restoration method that considers reference selection, transfer, and reconstruction (RefSTAR). In terms of selection, we construct a reference selection (RefSel) module. For training the RefSel module, we construct a RefSel-HQ dataset through a mask generation pipeline, which contains annotating masks for 10,000 ground truth-reference pairs. As for the transfer, due to the trivial solution in vanilla cross-attention operations, a feature fusion paradigm is designed to force the features from the reference to be integrated. Finally, we propose a reference image reconstruction mechanism that further ensures the presence of reference image features in the output image. The cycle consistency loss is also redesigned in conjunction with the mask. Extensive experiments on various backbone models demonstrate superior performance, showing better identity preservation ability and reference feature transfer quality. Source code, dataset, and pre-trained models are available at https://github.com/yinzhicun/RefSTAR.

Zheng Dong, Ke Xu, Yin Yang et al.

This paper proposes a novel location-aware deep-learning-based single image reflection removal method. Our network has a reflection detection module to regress a probabilistic reflection confidence map, taking multi-scale Laplacian features as inputs. This probabilistic map tells if a region is reflection-dominated or transmission-dominated, and it is used as a cue for the network to control the feature flow when predicting the reflection and transmission layers. We design our network as a recurrent network to progressively refine reflection removal results at each iteration. The novelty is that we leverage Laplacian kernel parameters to emphasize the boundaries of strong reflections. It is beneficial to strong reflection detection and substantially improves the quality of reflection removal results. Extensive experiments verify the superior performance of the proposed method over state-of-the-art approaches. Our code and the pre-trained model can be found at https://github.com/zdlarr/Location-aware-SIRR.

Zhanghan Ke, Jiayu Sun, Kaican Li et al.

Existing portrait matting methods either require auxiliary inputs that are costly to obtain or involve multiple stages that are computationally expensive, making them less suitable for real-time applications. In this work, we present a light-weight matting objective decomposition network (MODNet) for portrait matting in real-time with a single input image. The key idea behind our efficient design is by optimizing a series of sub-objectives simultaneously via explicit constraints. In addition, MODNet includes two novel techniques for improving model efficiency and robustness. First, an Efficient Atrous Spatial Pyramid Pooling (e-ASPP) module is introduced to fuse multi-scale features for semantic estimation. Second, a self-supervised sub-objectives consistency (SOC) strategy is proposed to adapt MODNet to real-world data to address the domain shift problem common to trimap-free methods. MODNet is easy to be trained in an end-to-end manner. It is much faster than contemporaneous methods and runs at 67 frames per second on a 1080Ti GPU. Experiments show that MODNet outperforms prior trimap-free methods by a large margin on both Adobe Matting Dataset and a carefully designed photographic portrait matting (PPM-100) benchmark proposed by us. Further, MODNet achieves remarkable results on daily photos and videos. Our code and models are available at https://github.com/ZHKKKe/MODNet, and the PPM-100 benchmark is released at https://github.com/ZHKKKe/PPM.

Zhanghan Ke, Di Qiu, Kaican Li et al.

We investigate the generalization of semi-supervised learning (SSL) to diverse pixel-wise tasks. Although SSL methods have achieved impressive results in image classification, the performances of applying them to pixel-wise tasks are unsatisfactory due to their need for dense outputs. In addition, existing pixel-wise SSL approaches are only suitable for certain tasks as they usually require to use task-specific properties. In this paper, we present a new SSL framework, named Guided Collaborative Training (GCT), for pixel-wise tasks, with two main technical contributions. First, GCT addresses the issues caused by the dense outputs through a novel flaw detector. Second, the modules in GCT learn from unlabeled data collaboratively through two newly proposed constraints that are independent of task-specific properties. As a result, GCT can be applied to a wide range of pixel-wise tasks without structural adaptation. Our extensive experiments on four challenging vision tasks, including semantic segmentation, real image denoising, portrait image matting, and night image enhancement, show that GCT outperforms state-of-the-art SSL methods by a large margin. Our code available at: https://github.com/ZHKKKe/PixelSSL.

Yuhao Liu, Zhanghan Ke, Fang Liu et al.

Diffusion models trained on large-scale datasets have achieved remarkable progress in image synthesis. However, due to the randomness in the diffusion process, they often struggle with handling diverse low-level tasks that require details preservation. To overcome this limitation, we present a new Diff-Plugin framework to enable a single pre-trained diffusion model to generate high-fidelity results across a variety of low-level tasks. Specifically, we first propose a lightweight Task-Plugin module with a dual branch design to provide task-specific priors, guiding the diffusion process in preserving image content. We then propose a Plugin-Selector that can automatically select different Task-Plugins based on the text instruction, allowing users to edit images by indicating multiple low-level tasks with natural language. We conduct extensive experiments on 8 low-level vision tasks. The results demonstrate the superiority of Diff-Plugin over existing methods, particularly in real-world scenarios. Our ablations further validate that Diff-Plugin is stable, schedulable, and supports robust training across different dataset sizes.

Yuhao Liu, Zhanghan Ke, Ke Xu et al.

Removing shadows requires an understanding of both lighting conditions and object textures in a scene. Existing methods typically learn pixel-level color mappings between shadow and non-shadow images, in which the joint modeling of lighting and object textures is implicit and inadequate. We observe that in a shadow region, the degradation degree of object textures depends on the local illumination, while simply enhancing the local illumination cannot fully recover the attenuated textures. Based on this observation, we propose to condition the restoration of attenuated textures on the corrected local lighting in the shadow region. Specifically, We first design a shadow-aware decomposition network to estimate the illumination and reflectance layers of shadow regions explicitly. We then propose a novel bilateral correction network to recast the lighting of shadow regions in the illumination layer via a novel local lighting correction module, and to restore the textures conditioned on the corrected illumination layer via a novel illumination-guided texture restoration module. We further annotate pixel-wise shadow masks for the public SRD dataset, which originally contains only image pairs. Experiments on three benchmarks show that our method outperforms existing state-of-the-art shadow removal methods.

Haoyuan Wang, Wenbo Hu, Lei Zhu et al.

Inverse rendering aims at recovering both geometry and materials of objects. It provides a more compatible reconstruction for conventional rendering engines, compared with the neural radiance fields (NeRFs). On the other hand, existing NeRF-based inverse rendering methods cannot handle glossy objects with local light interactions well, as they typically oversimplify the illumination as a 2D environmental map, which assumes infinite lights only. Observing the superiority of NeRFs in recovering radiance fields, we propose a novel 5D Neural Plenoptic Function (NeP) based on NeRFs and ray tracing, such that more accurate lighting-object interactions can be formulated via the rendering equation. We also design a material-aware cone sampling strategy to efficiently integrate lights inside the BRDF lobes with the help of pre-filtered radiance fields. Our method has two stages: the geometry of the target object and the pre-filtered environmental radiance fields are reconstructed in the first stage, and materials of the target object are estimated in the second stage with the proposed NeP and material-aware cone sampling strategy. Extensive experiments on the proposed real-world and synthetic datasets demonstrate that our method can reconstruct high-fidelity geometry/materials of challenging glossy objects with complex lighting interactions from nearby objects. Project webpage: https://whyy.site/paper/nep

Lei Zhu, Xinjiang Wang, Wayne Zhang et al.

A practical large language model (LLM) service may involve a long system prompt, which specifies the instructions, examples, and knowledge documents of the task and is reused across requests. However, the long system prompt causes throughput/latency bottlenecks as the cost of generating the next token grows w.r.t. the sequence length. This paper aims to improve the efficiency of LLM services that involve long system prompts. Our key observation is that handling these system prompts requires heavily redundant memory accesses in existing causal attention computation algorithms. Specifically, for batched requests, the cached hidden states (\ie, key-value pairs) of system prompts are transferred from off-chip DRAM to on-chip SRAM multiple times, each corresponding to an individual request. To eliminate such a redundancy, we propose RelayAttention, an attention algorithm that allows reading these hidden states from DRAM exactly once for a batch of input tokens. RelayAttention is a free lunch: it maintains the generation quality while requiring no model retraining, as it is based on a mathematical reformulation of causal attention. We have observed significant performance improvements to a production-level system, vLLM, through integration with RelayAttention. The improvements are even more profound with longer system prompts.

Zhenwei Wang, Tengfei Wang, Gerhard Hancke et al.

Real-world applications often require a large gallery of 3D assets that share a consistent theme. While remarkable advances have been made in general 3D content creation from text or image, synthesizing customized 3D assets following the shared theme of input 3D exemplars remains an open and challenging problem. In this work, we present ThemeStation, a novel approach for theme-aware 3D-to-3D generation. ThemeStation synthesizes customized 3D assets based on given few exemplars with two goals: 1) unity for generating 3D assets that thematically align with the given exemplars and 2) diversity for generating 3D assets with a high degree of variations. To this end, we design a two-stage framework that draws a concept image first, followed by a reference-informed 3D modeling stage. We propose a novel dual score distillation (DSD) loss to jointly leverage priors from both the input exemplars and the synthesized concept image. Extensive experiments and user studies confirm that ThemeStation surpasses prior works in producing diverse theme-aware 3D models with impressive quality. ThemeStation also enables various applications such as controllable 3D-to-3D generation.

Tianyu Huang, Wangguandong Zheng, Tengfei Wang et al.

Real-world applications like video gaming and virtual reality often demand the ability to model 3D scenes that users can explore along custom camera trajectories. While significant progress has been made in generating 3D objects from text or images, creating long-range, 3D-consistent, explorable 3D scenes remains a complex and challenging problem. In this work, we present Voyager, a novel video diffusion framework that generates world-consistent 3D point-cloud sequences from a single image with user-defined camera path. Unlike existing approaches, Voyager achieves end-to-end scene generation and reconstruction with inherent consistency across frames, eliminating the need for 3D reconstruction pipelines (e.g., structure-from-motion or multi-view stereo). Our method integrates three key components: 1) World-Consistent Video Diffusion: A unified architecture that jointly generates aligned RGB and depth video sequences, conditioned on existing world observation to ensure global coherence 2) Long-Range World Exploration: An efficient world cache with point culling and an auto-regressive inference with smooth video sampling for iterative scene extension with context-aware consistency, and 3) Scalable Data Engine: A video reconstruction pipeline that automates camera pose estimation and metric depth prediction for arbitrary videos, enabling large-scale, diverse training data curation without manual 3D annotations. Collectively, these designs result in a clear improvement over existing methods in visual quality and geometric accuracy, with versatile applications.

Jiaying Lin, Shuquan Ye, Rynson W. H. Lau

Multimodal Large Language Models (MLLMs) have achieved impressive results on various vision tasks, leveraging recent advancements in large language models. However, a critical question remains unaddressed: do MLLMs perceive visual information similarly to humans? Current benchmarks lack the ability to evaluate MLLMs from this perspective. To address this challenge, we introduce HVSBench, a large-scale benchmark designed to assess the alignment between MLLMs and the human visual system (HVS) on fundamental vision tasks that mirror human vision. HVSBench curated over 85K multimodal samples, spanning 13 categories and 5 fields in HVS, including Prominence, Subitizing, Prioritizing, Free-Viewing, and Searching. Extensive experiments demonstrate the effectiveness of our benchmark in providing a comprehensive evaluation of MLLMs. Specifically, we evaluate 13 MLLMs, revealing that even the best models show significant room for improvement, with most achieving only moderate results. Our experiments reveal that HVSBench presents a new and significant challenge for cutting-edge MLLMs. Diverse human participants attained strong performance, significantly outperforming MLLMs, which further underscores the benchmark's high quality. We believe that HVSBench will facilitate research on human-aligned and explainable MLLMs, marking a key step in understanding how MLLMs perceive and process visual information.

Tao Yan, Hao Huang, Yiwei Lu et al.

Glass surfaces are ubiquitous in daily life, typically appearing colorless, transparent, and lacking distinctive features. These characteristics make glass surface detection a challenging computer vision task. Existing glass surface detection methods always rely on boundary cues (e.g., window and door frames) or reflection cues to locate glass surfaces, but they fail to fully exploit the intrinsic properties of the glass itself for accurate localization. We observed that in most real-world scenes, the illumination intensity in front of the glass surface differs from that behind it, which results in variations in the reflections visible on the glass surface. Specifically, when standing on the brighter side of the glass and applying a flash towards the darker side, existing reflections on the glass surface tend to disappear. Conversely, while standing on the darker side and applying a flash towards the brighter side, distinct reflections will appear on the glass surface. Based on this phenomenon, we propose NFGlassNet, a novel method for glass surface detection that leverages the reflection dynamics present in flash/no-flash imagery. Specifically, we propose a Reflection Contrast Mining Module (RCMM) for extracting reflections, and a Reflection Guided Attention Module (RGAM) for fusing features from reflection and glass surface for accurate glass surface detection. For learning our network, we also construct a dataset consisting of 3.3K no-flash and flash image pairs captured from various scenes with corresponding ground truth annotations. Extensive experiments demonstrate that our method outperforms the state-of-the-art methods. Our code, model, and dataset will be available upon acceptance of the manuscript.

Zefan Qu, Zhenwei Wang, Haoyuan Wang et al.

Creating 3D assets that follow the texture and geometry style of existing ones is often desirable or even inevitable in practical applications like video gaming and virtual reality. While impressive progress has been made in generating 3D objects from text or images, creating style-controllable 3D assets remains a complex and challenging problem. In this work, we propose StyleSculptor, a novel training-free approach for generating style-guided 3D assets from a content image and one or more style images. Unlike previous works, StyleSculptor achieves style-guided 3D generation in a zero-shot manner, enabling fine-grained 3D style control that captures the texture, geometry, or both styles of user-provided style images. At the core of StyleSculptor is a novel Style Disentangled Attention (SD-Attn) module, which establishes a dynamic interaction between the input content image and style image for style-guided 3D asset generation via a cross-3D attention mechanism, enabling stable feature fusion and effective style-guided generation. To alleviate semantic content leakage, we also introduce a style-disentangled feature selection strategy within the SD-Attn module, which leverages the variance of 3D feature patches to disentangle style- and content-significant channels, allowing selective feature injection within the attention framework. With SD-Attn, the network can dynamically compute texture-, geometry-, or both-guided features to steer the 3D generation process. Built upon this, we further propose the Style Guided Control (SGC) mechanism, which enables exclusive geometry- or texture-only stylization, as well as adjustable style intensity control. Extensive experiments demonstrate that StyleSculptor outperforms existing baseline methods in producing high-fidelity 3D assets.

Dong Liang, Jinyuan Jia, Yuhao Liu et al.

While existing image-guided composition methods may help insert a foreground object onto a user-specified region of a background image, achieving natural blending inside the region with the rest of the image unchanged, we observe that these existing methods often struggle in synthesizing seamless interaction-aware compositions when the task involves human-object interactions. In this paper, we first propose HOComp, a novel approach for compositing a foreground object onto a human-centric background image, while ensuring harmonious interactions between the foreground object and the background person and their consistent appearances. Our approach includes two key designs: (1) MLLMs-driven Region-based Pose Guidance (MRPG), which utilizes MLLMs to identify the interaction region as well as the interaction type (e.g., holding and lefting) to provide coarse-to-fine constraints to the generated pose for the interaction while incorporating human pose landmarks to track action variations and enforcing fine-grained pose constraints; and (2) Detail-Consistent Appearance Preservation (DCAP), which unifies a shape-aware attention modulation mechanism, a multi-view appearance loss, and a background consistency loss to ensure consistent shapes/textures of the foreground and faithful reproduction of the background human. We then propose the first dataset, named Interaction-aware Human-Object Composition (IHOC), for the task. Experimental results on our dataset show that HOComp effectively generates harmonious human-object interactions with consistent appearances, and outperforms relevant methods qualitatively and quantitatively.

Yuhao Liu, Tengfei Wang, Fang Liu et al.

Recent advances in deep generative modeling have unlocked unprecedented opportunities for video synthesis. In real-world applications, however, users often seek tools to faithfully realize their creative editing intentions with precise and consistent control. Despite the progress achieved by existing methods, ensuring fine-grained alignment with user intentions remains an open and challenging problem. In this work, we present Shape-for-Motion, a novel framework that incorporates a 3D proxy for precise and consistent video editing. Shape-for-Motion achieves this by converting the target object in the input video to a time-consistent mesh, i.e., a 3D proxy, allowing edits to be performed directly on the proxy and then inferred back to the video frames. To simplify the editing process, we design a novel Dual-Propagation Strategy that allows users to perform edits on the 3D mesh of a single frame, and the edits are then automatically propagated to the 3D meshes of the other frames. The 3D meshes for different frames are further projected onto the 2D space to produce the edited geometry and texture renderings, which serve as inputs to a decoupled video diffusion model for generating edited results. Our framework supports various precise and physically-consistent manipulations across the video frames, including pose editing, rotation, scaling, translation, texture modification, and object composition. Our approach marks a key step toward high-quality, controllable video editing workflows. Extensive experiments demonstrate the superiority and effectiveness of our approach. Project page: https://shapeformotion.github.io/

Youjun Zhao, Jiaying Lin, Rynson W. H. Lau

Open-vocabulary 3D object detection (OV-3DOD) aims at localizing and classifying novel objects beyond closed sets. The recent success of vision-language models (VLMs) has demonstrated their remarkable capabilities to understand open vocabularies. Existing works that leverage VLMs for 3D object detection (3DOD) generally resort to representations that lose the rich scene context required for 3D perception. To address this problem, we propose in this paper a hierarchical framework, named HCMA, to simultaneously learn local object and global scene information for OV-3DOD. Specifically, we first design a Hierarchical Data Integration (HDI) approach to obtain coarse-to-fine 3D-image-text data, which is fed into a VLM to extract object-centric knowledge. To facilitate the association of feature hierarchies, we then propose an Interactive Cross-Modal Alignment (ICMA) strategy to establish effective intra-level and inter-level feature connections. To better align features across different levels, we further propose an Object-Focusing Context Adjustment (OFCA) module to refine multi-level features by emphasizing object-related features. Extensive experiments demonstrate that the proposed method outperforms SOTA methods on the existing OV-3DOD benchmarks. It also achieves promising OV-3DOD results even without any 3D annotations.

Tao Yan, Weijiang He, Chenglong Wang et al.

Since rainy weather always degrades image quality and poses significant challenges to most computer vision-based intelligent systems, image de-raining has been a hot research topic. Fortunately, in a rainy light field (LF) image, background obscured by rain streaks in one sub-view may be visible in the other sub-views, and implicit depth information and recorded 4D structural information may benefit rain streak detection and removal. However, existing LF image rain removal methods either do not fully exploit the global correlations of 4D LF data or only utilize partial sub-views, resulting in sub-optimal rain removal performance and no-equally good quality for all de-rained sub-views. In this paper, we propose an efficient network, called MDeRainNet, for rain streak removal from LF images. The proposed network adopts a multi-scale encoder-decoder architecture, which directly works on Macro-pixel images (MPIs) to improve the rain removal performance. To fully model the global correlation between the spatial and the angular information, we propose an Extended Spatial-Angular Interaction (ESAI) module to merge them, in which a simple and effective Transformer-based Spatial-Angular Interaction Attention (SAIA) block is also proposed for modeling long-range geometric correlations and making full use of the angular information. Furthermore, to improve the generalization performance of our network on real-world rainy scenes, we propose a novel semi-supervised learning framework for our MDeRainNet, which utilizes multi-level KL loss to bridge the domain gap between features of synthetic and real-world rain streaks and introduces colored-residue image guided contrastive regularization to reconstruct rain-free images. Extensive experiments conducted on synthetic and real-world LFIs demonstrate that our method outperforms the state-of-the-art methods both quantitatively and qualitatively.

Tianyu Huang, Haoze Zhang, Yihan Zeng et al.

Dynamic 3D interaction has been attracting a lot of attention recently. However, creating such 4D content remains challenging. One solution is to animate 3D scenes with physics-based simulation, which requires manually assigning precise physical properties to the object or the simulated results would become unnatural. Another solution is to learn the deformation of 3D objects with the distillation of video generative models, which, however, tends to produce 3D videos with small and discontinuous motions due to the inappropriate extraction and application of physics priors. In this work, to combine the strengths and complementing shortcomings of the above two solutions, we propose to learn the physical properties of a material field with video diffusion priors, and then utilize a physics-based Material-Point-Method (MPM) simulator to generate 4D content with realistic motions. In particular, we propose motion distillation sampling to emphasize video motion information during distillation. In addition, to facilitate the optimization, we further propose a KAN-based material field with frame boosting. Experimental results demonstrate that our method enjoys more realistic motions than state-of-the-arts do.

Jiaying Lin, Huankang Guan, Rynson W. H. Lau

Salient Object Ranking (SOR) involves ranking the degree of saliency of multiple salient objects in an input image. Most recently, a method is proposed for ranking salient objects in an input video based on a predicted fixation map. It relies solely on the density of the fixations within the salient objects to infer their saliency ranks, which is incompatible with human perception of saliency ranking. In this work, we propose to explicitly learn the spatial and temporal relations between different salient objects to produce the saliency ranks. To this end, we propose an end-to-end method for video salient object ranking (VSOR), with two novel modules: an intra-frame adaptive relation (IAR) module to learn the spatial relation among the salient objects in the same frame locally and globally, and an inter-frame dynamic relation (IDR) module to model the temporal relation of saliency across different frames. In addition, to address the limited video types (just sports and movies) and scene diversity in the existing VSOR dataset, we propose a new dataset that covers different video types and diverse scenes on a large scale. Experimental results demonstrate that our method outperforms state-of-the-art methods in relevant fields. We will make the source code and our proposed dataset available.

Zheng Dong, Ke Xu, Ziheng Duan et al.

Although PIFu-based 3D human reconstruction methods are popular, the quality of recovered details is still unsatisfactory. In a sparse (e.g., 3 RGBD sensors) capture setting, the depth noise is typically amplified in the PIFu representation, resulting in flat facial surfaces and geometry-fallible bodies. In this paper, we propose a novel geometry-aware two-scale PIFu for 3D human reconstruction from sparse, noisy inputs. Our key idea is to exploit the complementary properties of depth denoising and 3D reconstruction, for learning a two-scale PIFu representation to reconstruct high-frequency facial details and consistent bodies separately. To this end, we first formulate depth denoising and 3D reconstruction as a multi-task learning problem. The depth denoising process enriches the local geometry information of the reconstruction features, while the reconstruction process enhances depth denoising with global topology information. We then propose to learn the two-scale PIFu representation using two MLPs based on the denoised depth and geometry-aware features. Extensive experiments demonstrate the effectiveness of our approach in reconstructing facial details and bodies of different poses and its superiority over state-of-the-art methods.

Xin Tian, Ke Xu, Xin Yang et al.

Existing salient instance detection (SID) methods typically learn from pixel-level annotated datasets. In this paper, we present the first weakly-supervised approach to the SID problem. Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization. However, it is non-trivial to use only class labels to learn instance-aware saliency information, as salient instances with high semantic affinities may not be easily separated by the labels. As the subitizing information provides an instant judgement on the number of salient items, it is naturally related to detecting salient instances and may help separate instances of the same class while grouping different parts of the same instance. Inspired by this observation, we propose to use class and subitizing labels as weak supervision for the SID problem. We propose a novel weakly-supervised network with three branches: a Saliency Detection Branch leveraging class consistency information to locate candidate objects; a Boundary Detection Branch exploiting class discrepancy information to delineate object boundaries; and a Centroid Detection Branch using subitizing information to detect salient instance centroids. This complementary information is then fused to produce a salient instance map. To facilitate the learning process, we further propose a progressive training scheme to reduce label noise and the corresponding noise learned by the model, via reciprocating the model with progressive salient instance prediction and model refreshing. Our extensive evaluations show that the proposed method plays favorably against carefully designed baseline methods adapted from related tasks.

Jiayu Sun, Zhanghan Ke, Lihe Zhang et al.

To address the challenging portrait video matting problem more precisely, existing works typically apply some matting priors that require additional user efforts to obtain, such as annotated trimaps or background images. In this work, we observe that instead of asking the user to explicitly provide a background image, we may recover it from the input video itself. To this end, we first propose a novel background restoration module (BRM) to recover the background image dynamically from the input video. BRM is extremely lightweight and can be easily integrated into existing matting models. By combining BRM with a recent image matting model, MODNet, we then present MODNet-V for portrait video matting. Benefited from the strong background prior provided by BRM, MODNet-V has only 1/3 of the parameters of MODNet but achieves comparable or even better performances. Our design allows MODNet-V to be trained in an end-to-end manner on a single NVIDIA 3090 GPU. Finally, we introduce a new patch refinement module (PRM) to adapt MODNet-V for high-resolution videos while keeping MODNet-V lightweight and fast.

Xin Yang, Yu Qiao, Shaozhe Chen et al.

Image matting is an ill-posed problem that usually requires additional user input, such as trimaps or scribbles. Drawing a fne trimap requires a large amount of user effort, while using scribbles can hardly obtain satisfactory alpha mattes for non-professional users. Some recent deep learning-based matting networks rely on large-scale composite datasets for training to improve performance, resulting in the occasional appearance of obvious artifacts when processing natural images. In this article, we explore the intrinsic relationship between user input and alpha mattes and strike a balance between user effort and the quality of alpha mattes. In particular, we propose an interactive framework, referred to as smart scribbles, to guide users to draw few scribbles on the input images to produce high-quality alpha mattes. It frst infers the most informative regions of an image for drawing scribbles to indicate different categories (foreground, background, or unknown) and then spreads these scribbles (i.e., the category labels) to the rest of the image via our well-designed two-phase propagation. Both neighboring low-level afnities and high-level semantic features are considered during the propagation process. Our method can be optimized without large-scale matting datasets and exhibits more universality in real situations. Extensive experiments demonstrate that smart scribbles can produce more accurate alpha mattes with reduced additional input, compared to the state-of-the-art matting methods.

Xin Yang, Zongliang Ma, Letian Yu et al.

In this paper, we propose a fully automatic system for generating comic books from videos without any human intervention. Given an input video along with its subtitles, our approach first extracts informative keyframes by analyzing the subtitles, and stylizes keyframes into comic-style images. Then, we propose a novel automatic multi-page layout framework, which can allocate the images across multiple pages and synthesize visually interesting layouts based on the rich semantics of the images (e.g., importance and inter-image relation). Finally, as opposed to using the same type of balloon as in previous works, we propose an emotion-aware balloon generation method to create different types of word balloons by analyzing the emotion of subtitles and audios. Our method is able to vary balloon shapes and word sizes in balloons in response to different emotions, leading to more enriched reading experience. Once the balloons are generated, they are placed adjacent to their corresponding speakers via speaker detection. Our results show that our method, without requiring any user inputs, can generate high-quality comic pages with visually rich layouts and balloons. Our user studies also demonstrate that users prefer our generated results over those by state-of-the-art comic generation systems.

Xiaoyang Zheng, Xin Tan, Jie Zhou et al.

Salient object detection aims at detecting the most visually distinct objects and producing the corresponding masks. As the cost of pixel-level annotations is high, image tags are usually used as weak supervisions. However, an image tag can only be used to annotate one class of objects. In this paper, we introduce saliency subitizing as the weak supervision since it is class-agnostic. This allows the supervision to be aligned with the property of saliency detection, where the salient objects of an image could be from more than one class. To this end, we propose a model with two modules, Saliency Subitizing Module (SSM) and Saliency Updating Module (SUM). While SSM learns to generate the initial saliency masks using the subitizing information, without the need for any unsupervised methods or some random seeds, SUM helps iteratively refine the generated saliency masks. We conduct extensive experiments on five benchmark datasets. The experimental results show that our method outperforms other weakly-supervised methods and even performs comparably to some fully-supervised methods.

Xin Tian, Ke Xu, Xin Yang et al.

Existing salient instance detection (SID) methods typically learn from pixel-level annotated datasets. In this paper, we present the first weakly-supervised approach to the SID problem. Although weak supervision has been considered in general saliency detection, it is mainly based on using class labels for object localization. However, it is non-trivial to use only class labels to learn instance-aware saliency information, as salient instances with high semantic affinities may not be easily separated by the labels. We note that subitizing information provides an instant judgement on the number of salient items, which naturally relates to detecting salient instances and may help separate instances of the same class while grouping different parts of the same instance. Inspired by this insight, we propose to use class and subitizing labels as weak supervision for the SID problem. We propose a novel weakly-supervised network with three branches: a Saliency Detection Branch leveraging class consistency information to locate candidate objects; a Boundary Detection Branch exploiting class discrepancy information to delineate object boundaries; and a Centroid Detection Branch using subitizing information to detect salient instance centroids. This complementary information is further fused to produce salient instance maps. We conduct extensive experiments to demonstrate that the proposed method plays favorably against carefully designed baseline methods adapted from related tasks.

Yuzhen Niu, Jianbin Wu, Wenxi Liu et al.

Synthesizing high dynamic range (HDR) images from multiple low-dynamic range (LDR) exposures in dynamic scenes is challenging. There are two major problems caused by the large motions of foreground objects. One is the severe misalignment among the LDR images. The other is the missing content due to the over-/under-saturated regions caused by the moving objects, which may not be easily compensated for by the multiple LDR exposures. Thus, it requires the HDR generation model to be able to properly fuse the LDR images and restore the missing details without introducing artifacts. To address these two problems, we propose in this paper a novel GAN-based model, HDR-GAN, for synthesizing HDR images from multi-exposed LDR images. To our best knowledge, this work is the first GAN-based approach for fusing multi-exposed LDR images for HDR reconstruction. By incorporating adversarial learning, our method is able to produce faithful information in the regions with missing content. In addition, we also propose a novel generator network, with a reference-based residual merging block for aligning large object motions in the feature domain, and a deep HDR supervision scheme for eliminating artifacts of the reconstructed HDR images. Experimental results demonstrate that our model achieves state-of-the-art reconstruction performance over the prior HDR methods on diverse scenes.

Nanxuan Zhao, Zhirong Wu, Rynson W. H. Lau et al.

Contrastive visual pretraining based on the instance discrimination pretext task has made significant progress. Notably, recent work on unsupervised pretraining has shown to surpass the supervised counterpart for finetuning downstream applications such as object detection and segmentation. It comes as a surprise that image annotations would be better left unused for transfer learning. In this work, we investigate the following problems: What makes instance discrimination pretraining good for transfer learning? What knowledge is actually learned and transferred from these models? From this understanding of instance discrimination, how can we better exploit human annotation labels for pretraining? Our findings are threefold. First, what truly matters for the transfer is low-level and mid-level representations, not high-level representations. Second, the intra-category invariance enforced by the traditional supervised model weakens transferability by increasing task misalignment. Finally, supervised pretraining can be strengthened by following an exemplar-based approach without explicit constraints among the instances within the same category.