Wangmeng Zuo

Haoliang Meng, Xiaopeng Hong, Yabin Wang et al.

This paper constructs the first benchmark on semi-supervised multi-modal crowd counting. To lay the foundation for this unexplored task, we first formulate the semi-supervised multi-modal setting and a standardized protocol that specifies the labeled-unlabeled data partition across different labeled ratios. Next, to establish solid reference points, we carefully tailor a diverse set of representative baselines, including existing fully supervised multi-modal methods and semi-supervised single-modal methods. Then, we carefully evaluate their performance under our proposed benchmark. Codes and the data partition will be released on https://github.com/HenryCilence/Semi-supervised-Multimodal-Crowd-Counting.

Yuxiang Wei, Yabo Zhang, Zhilong Ji et al. · microsoft-research

In addition to the unprecedented ability in imaginary creation, large text-to-image models are expected to take customized concepts in image generation. Existing works generally learn such concepts in an optimization-based manner, yet bringing excessive computation or memory burden. In this paper, we instead propose a learning-based encoder, which consists of a global and a local mapping networks for fast and accurate customized text-to-image generation. In specific, the global mapping network projects the hierarchical features of a given image into multiple new words in the textual word embedding space, i.e., one primary word for well-editable concept and other auxiliary words to exclude irrelevant disturbances (e.g., background). In the meantime, a local mapping network injects the encoded patch features into cross attention layers to provide omitted details, without sacrificing the editability of primary concepts. We compare our method with existing optimization-based approaches on a variety of user-defined concepts, and demonstrate that our method enables high-fidelity inversion and more robust editability with a significantly faster encoding process. Our code is publicly available at https://github.com/csyxwei/ELITE.

Zehao Wang, Guanglei Yang, Yihan Zeng et al.

Federated fine-tuning of foundation models with Low-Rank Adaptation (LoRA) provides an efficient solution for reducing communication and computation costs while preserving data locality. However, the direct combination of FedAvg and LoRA suffers from three key issues: limited update space, which restricts the model's effective learning capacity; inter-round state mismatch, which disrupts cross-round local optimization continuity; and a client-agnostic starting state, which slows local convergence on clients. Although recent methods mitigate the limited update space issue by merging LoRA updates into the backbone across communication rounds, inter-round state mismatch and the client-agnostic starting state remain insufficiently addressed. To address these issues, we propose FedSmoothLoRA, a federated LoRA tuning framework that preserves the enlarged update space, improves cross-round local optimization continuity, and provides a client-aware starting state for local training. At each communication round, FedSmoothLoRA constructs the local LoRA initialization using two matrices: a Round-Matching matrix that preserves cross-round local state continuity, and a Gradient-Aligned matrix that provides client-specific optimization guidance from gradient signals estimated on local data. Together, these designs enable smoother and faster convergence. Extensive experiments on image classification and natural language generation tasks demonstrate that FedSmoothLoRA consistently outperforms existing federated LoRA tuning methods. Code: https://github.com/wangzehao0704/FedSmoothLoRA

Ren Yang, Radu Timofte, Meisong Zheng et al. · tencent-ai

This paper reviews the NTIRE 2022 Challenge on Super-Resolution and Quality Enhancement of Compressed Video. In this challenge, we proposed the LDV 2.0 dataset, which includes the LDV dataset (240 videos) and 95 additional videos. This challenge includes three tracks. Track 1 aims at enhancing the videos compressed by HEVC at a fixed QP. Track 2 and Track 3 target both the super-resolution and quality enhancement of HEVC compressed video. They require x2 and x4 super-resolution, respectively. The three tracks totally attract more than 600 registrations. In the test phase, 8 teams, 8 teams and 12 teams submitted the final results to Tracks 1, 2 and 3, respectively. The proposed methods and solutions gauge the state-of-the-art of super-resolution and quality enhancement of compressed video. The proposed LDV 2.0 dataset is available at https://github.com/RenYang-home/LDV_dataset. The homepage of this challenge (including open-sourced codes) is at https://github.com/RenYang-home/NTIRE22_VEnh_SR.

Qiying Yu, Jieming Lou, Xianyuan Zhan et al. · tsinghua

Contrastive learning (CL) has recently been applied to adversarial learning tasks. Such practice considers adversarial samples as additional positive views of an instance, and by maximizing their agreements with each other, yields better adversarial robustness. However, this mechanism can be potentially flawed, since adversarial perturbations may cause instance-level identity confusion, which can impede CL performance by pulling together different instances with separate identities. To address this issue, we propose to treat adversarial samples unequally when contrasted, with an asymmetric InfoNCE objective ($A-InfoNCE$) that allows discriminating considerations of adversarial samples. Specifically, adversaries are viewed as inferior positives that induce weaker learning signals, or as hard negatives exhibiting higher contrast to other negative samples. In the asymmetric fashion, the adverse impacts of conflicting objectives between CL and adversarial learning can be effectively mitigated. Experiments show that our approach consistently outperforms existing Adversarial CL methods across different finetuning schemes without additional computational cost. The proposed A-InfoNCE is also a generic form that can be readily extended to other CL methods. Code is available at https://github.com/yqy2001/A-InfoNCE.

Chunwei Tian, Menghua Zheng, Wangmeng Zuo et al.

Deep convolutional neural networks (CNNs) are used for image denoising via automatically mining accurate structure information. However, most of existing CNNs depend on enlarging depth of designed networks to obtain better denoising performance, which may cause training difficulty. In this paper, we propose a multi-stage image denoising CNN with the wavelet transform (MWDCNN) via three stages, i.e., a dynamic convolutional block (DCB), two cascaded wavelet transform and enhancement blocks (WEBs) and a residual block (RB). DCB uses a dynamic convolution to dynamically adjust parameters of several convolutions for making a tradeoff between denoising performance and computational costs. WEB uses a combination of signal processing technique (i.e., wavelet transformation) and discriminative learning to suppress noise for recovering more detailed information in image denoising. To further remove redundant features, RB is used to refine obtained features for improving denoising effects and reconstruct clean images via improved residual dense architectures. Experimental results show that the proposed MWDCNN outperforms some popular denoising methods in terms of quantitative and qualitative analysis. Codes are available at https://github.com/hellloxiaotian/MWDCNN.

Chunwei Tian, Yixuan Yuan, Shichao Zhang et al.

CNNs with strong learning abilities are widely chosen to resolve super-resolution problem. However, CNNs depend on deeper network architectures to improve performance of image super-resolution, which may increase computational cost in general. In this paper, we present an enhanced super-resolution group CNN (ESRGCNN) with a shallow architecture by fully fusing deep and wide channel features to extract more accurate low-frequency information in terms of correlations of different channels in single image super-resolution (SISR). Also, a signal enhancement operation in the ESRGCNN is useful to inherit more long-distance contextual information for resolving long-term dependency. An adaptive up-sampling operation is gathered into a CNN to obtain an image super-resolution model with low-resolution images of different sizes. Extensive experiments report that our ESRGCNN surpasses the state-of-the-arts in terms of SISR performance, complexity, execution speed, image quality evaluation and visual effect in SISR. Code is found at https://github.com/hellloxiaotian/ESRGCNN.

Zixian Guo, Bowen Dong, Zhilong Ji et al.

Prompt tuning has been employed as an efficient way to adapt large vision-language pre-trained models (e.g. CLIP) to various downstream tasks in data-limited or label-limited settings. Nonetheless, visual data (e.g., images) is by default prerequisite for learning prompts in existing methods. In this work, we advocate that the effectiveness of image-text contrastive learning in aligning the two modalities (for training CLIP) further makes it feasible to treat texts as images for prompt tuning and introduce TaI prompting. In contrast to the visual data, text descriptions are easy to collect, and their class labels can be directly derived. Particularly, we apply TaI prompting to multi-label image recognition, where sentences in the wild serve as alternatives to images for prompt tuning. Moreover, with TaI, double-grained prompt tuning (TaI-DPT) is further presented to extract both coarse-grained and fine-grained embeddings for enhancing the multi-label recognition performance. Experimental results show that our proposed TaI-DPT outperforms zero-shot CLIP by a large margin on multiple benchmarks, e.g., MS-COCO, VOC2007, and NUS-WIDE, while it can be combined with existing methods of prompting from images to improve recognition performance further. Code is released at https://github.com/guozix/TaI-DPT.

Chunwei Tian, Yanning Zhang, Wangmeng Zuo et al.

Convolutional neural networks (CNNs) have obtained remarkable performance via deep architectures. However, these CNNs often achieve poor robustness for image super-resolution (SR) under complex scenes. In this paper, we present a heterogeneous group SR CNN (HGSRCNN) via leveraging structure information of different types to obtain a high-quality image. Specifically, each heterogeneous group block (HGB) of HGSRCNN uses a heterogeneous architecture containing a symmetric group convolutional block and a complementary convolutional block in a parallel way to enhance internal and external relations of different channels for facilitating richer low-frequency structure information of different types. To prevent appearance of obtained redundant features, a refinement block with signal enhancements in a serial way is designed to filter useless information. To prevent loss of original information, a multi-level enhancement mechanism guides a CNN to achieve a symmetric architecture for promoting expressive ability of HGSRCNN. Besides, a parallel up-sampling mechanism is developed to train a blind SR model. Extensive experiments illustrate that the proposed HGSRCNN has obtained excellent SR performance in terms of both quantitative and qualitative analysis. Codes can be accessed at https://github.com/hellloxiaotian/HGSRCNN.

Zhaohui Zheng, Rongguang Ye, Qibin Hou et al.

Previous knowledge distillation (KD) methods for object detection mostly focus on feature imitation instead of mimicking the prediction logits due to its inefficiency in distilling the localization information. In this paper, we investigate whether logit mimicking always lags behind feature imitation. Towards this goal, we first present a novel localization distillation (LD) method which can efficiently transfer the localization knowledge from the teacher to the student. Second, we introduce the concept of valuable localization region that can aid to selectively distill the classification and localization knowledge for a certain region. Combining these two new components, for the first time, we show that logit mimicking can outperform feature imitation and the absence of localization distillation is a critical reason for why logit mimicking underperforms for years. The thorough studies exhibit the great potential of logit mimicking that can significantly alleviate the localization ambiguity, learn robust feature representation, and ease the training difficulty in the early stage. We also provide the theoretical connection between the proposed LD and the classification KD, that they share the equivalent optimization effect. Our distillation scheme is simple as well as effective and can be easily applied to both dense horizontal object detectors and rotated object detectors. Extensive experiments on the MS COCO, PASCAL VOC, and DOTA benchmarks demonstrate that our method can achieve considerable AP improvement without any sacrifice on the inference speed. Our source code and pretrained models are publicly available at https://github.com/HikariTJU/LD.

Junyi Li, Zhilu Zhang, Xiaoyu Liu et al.

Significant progress has been made in self-supervised image denoising (SSID) in the recent few years. However, most methods focus on dealing with spatially independent noise, and they have little practicality on real-world sRGB images with spatially correlated noise. Although pixel-shuffle downsampling has been suggested for breaking the noise correlation, it breaks the original information of images, which limits the denoising performance. In this paper, we propose a novel perspective to solve this problem, i.e., seeking for spatially adaptive supervision for real-world sRGB image denoising. Specifically, we take into account the respective characteristics of flat and textured regions in noisy images, and construct supervisions for them separately. For flat areas, the supervision can be safely derived from non-adjacent pixels, which are much far from the current pixel for excluding the influence of the noise-correlated ones. And we extend the blind-spot network to a blind-neighborhood network (BNN) for providing supervision on flat areas. For textured regions, the supervision has to be closely related to the content of adjacent pixels. And we present a locally aware network (LAN) to meet the requirement, while LAN itself is selectively supervised with the output of BNN. Combining these two supervisions, a denoising network (e.g., U-Net) can be well-trained. Extensive experiments show that our method performs favorably against state-of-the-art SSID methods on real-world sRGB photographs. The code is available at https://github.com/nagejacob/SpatiallyAdaptiveSSID.

Qizhang Li, Yiwen Guo, Wangmeng Zuo et al.

The transferability of adversarial examples across deep neural networks (DNNs) is the crux of many black-box attacks. Many prior efforts have been devoted to improving the transferability via increasing the diversity in inputs of some substitute models. In this paper, by contrast, we opt for the diversity in substitute models and advocate to attack a Bayesian model for achieving desirable transferability. Deriving from the Bayesian formulation, we develop a principled strategy for possible finetuning, which can be combined with many off-the-shelf Gaussian posterior approximations over DNN parameters. Extensive experiments have been conducted to verify the effectiveness of our method, on common benchmark datasets, and the results demonstrate that our method outperforms recent state-of-the-arts by large margins (roughly 19% absolute increase in average attack success rate on ImageNet), and, by combining with these recent methods, further performance gain can be obtained. Our code: https://github.com/qizhangli/MoreBayesian-attack.

Qizhang Li, Yiwen Guo, Wangmeng Zuo et al.

Intermediate-level attacks that attempt to perturb feature representations following an adversarial direction drastically have shown favorable performance in crafting transferable adversarial examples. Existing methods in this category are normally formulated with two separate stages, where a directional guide is required to be determined at first and the scalar projection of the intermediate-level perturbation onto the directional guide is enlarged thereafter. The obtained perturbation deviates from the guide inevitably in the feature space, and it is revealed in this paper that such a deviation may lead to sub-optimal attack. To address this issue, we develop a novel intermediate-level method that crafts adversarial examples within a single stage of optimization. In particular, the proposed method, named intermediate-level perturbation decay (ILPD), encourages the intermediate-level perturbation to be in an effective adversarial direction and to possess a great magnitude simultaneously. In-depth discussion verifies the effectiveness of our method. Experimental results show that it outperforms state-of-the-arts by large margins in attacking various victim models on ImageNet (+10.07% on average) and CIFAR-10 (+3.88% on average). Our code is at https://github.com/qizhangli/ILPD-attack.

Minheng Ni, Yabo Zhang, Kailai Feng et al.

Zero-shot referring image segmentation is a challenging task because it aims to find an instance segmentation mask based on the given referring descriptions, without training on this type of paired data. Current zero-shot methods mainly focus on using pre-trained discriminative models (e.g., CLIP). However, we have observed that generative models (e.g., Stable Diffusion) have potentially understood the relationships between various visual elements and text descriptions, which are rarely investigated in this task. In this work, we introduce a novel Referring Diffusional segmentor (Ref-Diff) for this task, which leverages the fine-grained multi-modal information from generative models. We demonstrate that without a proposal generator, a generative model alone can achieve comparable performance to existing SOTA weakly-supervised models. When we combine both generative and discriminative models, our Ref-Diff outperforms these competing methods by a significant margin. This indicates that generative models are also beneficial for this task and can complement discriminative models for better referring segmentation. Our code is publicly available at https://github.com/kodenii/Ref-Diff.

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.

Junyi Li, Xiaohe Wu, Zhenxing Niu et al.

While significant progress has been made in deep video denoising, it remains very challenging for exploiting historical and future frames. Bidirectional recurrent networks (BiRNN) have exhibited appealing performance in several video restoration tasks. However, BiRNN is intrinsically offline because it uses backward recurrent modules to propagate from the last to current frames, which causes high latency and large memory consumption. To address the offline issue of BiRNN, we present a novel recurrent network consisting of forward and look-ahead recurrent modules for unidirectional video denoising. Particularly, look-ahead module is an elaborate forward module for leveraging information from near-future frames. When denoising the current frame, the hidden features by forward and look-ahead recurrent modules are combined, thereby making it feasible to exploit both historical and near-future frames. Due to the scene motion between non-neighboring frames, border pixels missing may occur when warping look-ahead feature from near-future frame to current frame, which can be largely alleviated by incorporating forward warping and proposed border enlargement. Experiments show that our method achieves state-of-the-art performance with constant latency and memory consumption. Code is avaliable at https://github.com/nagejacob/FloRNN.

Zhilu Zhang, Ruohao Wang, Hongzhi Zhang et al.

In this paper, we consider two challenging issues in reference-based super-resolution (RefSR), (i) how to choose a proper reference image, and (ii) how to learn real-world RefSR in a self-supervised manner. Particularly, we present a novel self-supervised learning approach for real-world image SR from observations at dual camera zooms (SelfDZSR). Considering the popularity of multiple cameras in modern smartphones, the more zoomed (telephoto) image can be naturally leveraged as the reference to guide the SR of the lesser zoomed (short-focus) image. Furthermore, SelfDZSR learns a deep network to obtain the SR result of short-focus image to have the same resolution as the telephoto image. For this purpose, we take the telephoto image instead of an additional high-resolution image as the supervision information and select a center patch from it as the reference to super-resolve the corresponding short-focus image patch. To mitigate the effect of the misalignment between short-focus low-resolution (LR) image and telephoto ground-truth (GT) image, we design an auxiliary-LR generator and map the GT to an auxiliary-LR while keeping the spatial position unchanged. Then the auxiliary-LR can be utilized to deform the LR features by the proposed adaptive spatial transformer networks (AdaSTN), and match the Ref features to GT. During testing, SelfDZSR can be directly deployed to super-solve the whole short-focus image with the reference of telephoto image. Experiments show that our method achieves better quantitative and qualitative performance against state-of-the-arts. Codes are available at https://github.com/cszhilu1998/SelfDZSR.

Yabo Zhang, Mingshuai Yao, Yuxiang Wei et al.

One-shot generative domain adaption aims to transfer a pre-trained generator on one domain to a new domain using one reference image only. However, it remains very challenging for the adapted generator (i) to generate diverse images inherited from the pre-trained generator while (ii) faithfully acquiring the domain-specific attributes and styles of the reference image. In this paper, we present a novel one-shot generative domain adaption method, i.e., DiFa, for diverse generation and faithful adaptation. For global-level adaptation, we leverage the difference between the CLIP embedding of reference image and the mean embedding of source images to constrain the target generator. For local-level adaptation, we introduce an attentive style loss which aligns each intermediate token of adapted image with its corresponding token of the reference image. To facilitate diverse generation, selective cross-domain consistency is introduced to select and retain the domain-sharing attributes in the editing latent $\mathcal{W}+$ space to inherit the diversity of pre-trained generator. Extensive experiments show that our method outperforms the state-of-the-arts both quantitatively and qualitatively, especially for the cases of large domain gaps. Moreover, our DiFa can easily be extended to zero-shot generative domain adaption with appealing results. Code is available at https://github.com/1170300521/DiFa.

Xiaoming Li, Chaofeng Chen, Xianhui Lin et al.

How to design proper training pairs is critical for super-resolving real-world low-quality (LQ) images, which suffers from the difficulties in either acquiring paired ground-truth high-quality (HQ) images or synthesizing photo-realistic degraded LQ observations. Recent works mainly focus on modeling the degradation with handcrafted or estimated degradation parameters, which are however incapable to model complicated real-world degradation types, resulting in limited quality improvement. Notably, LQ face images, which may have the same degradation process as natural images, can be robustly restored with photo-realistic textures by exploiting their strong structural priors. This motivates us to use the real-world LQ face images and their restored HQ counterparts to model the complex real-world degradation (namely ReDegNet), and then transfer it to HQ natural images to synthesize their realistic LQ counterparts. By taking these paired HQ-LQ face images as inputs to explicitly predict the degradation-aware and content-independent representations, we could control the degraded image generation, and subsequently transfer these degradation representations from face to natural images to synthesize the degraded LQ natural images. Experiments show that our ReDegNet can well learn the real degradation process from face images. The restoration network trained with our synthetic pairs performs favorably against SOTAs. More importantly, our method provides a new way to handle the real-world complex scenarios by learning their degradation representations from the facial portions, which can be used to significantly improve the quality of non-facial areas. The source code is available at https://github.com/csxmli2016/ReDegNet.

Zitong Huang, Yiping Bao, Bowen Dong et al.

Weakly-supervised object detection (WSOD) aims to train an object detector only requiring the image-level annotations. Recently, some works have managed to select the accurate boxes generated from a well-trained WSOD network to supervise a semi-supervised detection framework for better performance. However, these approaches simply divide the training set into labeled and unlabeled sets according to the image-level criteria, such that sufficient mislabeled or wrongly localized box predictions are chosen as pseudo ground-truths, resulting in a sub-optimal solution of detection performance. To overcome this issue, we propose a novel WSOD framework with a new paradigm that switches from weak supervision to noisy supervision (W2N). Generally, with given pseudo ground-truths generated from the well-trained WSOD network, we propose a two-module iterative training algorithm to refine pseudo labels and supervise better object detector progressively. In the localization adaptation module, we propose a regularization loss to reduce the proportion of discriminative parts in original pseudo ground-truths, obtaining better pseudo ground-truths for further training. In the semi-supervised module, we propose a two tasks instance-level split method to select high-quality labels for training a semi-supervised detector. Experimental results on different benchmarks verify the effectiveness of W2N, and our W2N outperforms all existing pure WSOD methods and transfer learning methods. Our code is publicly available at https://github.com/1170300714/w2n_wsod.

Zhilu Zhang, Rongjian Xu, Ming Liu et al.

When taking photos under an environment with insufficient light, the exposure time and the sensor gain usually require to be carefully chosen to obtain images with satisfying visual quality. For example, the images with high ISO usually have inescapable noise, while the long-exposure ones may be blurry due to camera shake or object motion. Existing solutions generally suggest to seek a balance between noise and blur, and learn denoising or deblurring models under either full- or self-supervision. However, the real-world training pairs are difficult to collect, and the self-supervised methods merely rely on blurry or noisy images are limited in performance. In this work, we tackle this problem by jointly leveraging the short-exposure noisy image and the long-exposure blurry image for better image restoration. Such setting is practically feasible due to that short-exposure and long-exposure images can be either acquired by two individual cameras or synthesized by a long burst of images. Moreover, the short-exposure images are hardly blurry, and the long-exposure ones have negligible noise. Their complementarity makes it feasible to learn restoration model in a self-supervised manner. Specifically, the noisy images can be used as the supervision information for deblurring, while the sharp areas in the blurry images can be utilized as the auxiliary supervision information for self-supervised denoising. By learning in a collaborative manner, the deblurring and denoising tasks in our method can benefit each other. Experiments on synthetic and real-world images show the effectiveness and practicality of the proposed method. Codes are available at https://github.com/cszhilu1998/SelfIR.

Qizhang Li, Yiwen Guo, Wangmeng Zuo et al.

The vulnerability of deep neural networks (DNNs) to adversarial examples has attracted great attention in the machine learning community. The problem is related to non-flatness and non-smoothness of normally obtained loss landscapes. Training augmented with adversarial examples (a.k.a., adversarial training) is considered as an effective remedy. In this paper, we highlight that some collaborative examples, nearly perceptually indistinguishable from both adversarial and benign examples yet show extremely lower prediction loss, can be utilized to enhance adversarial training. A novel method is therefore proposed to achieve new state-of-the-arts in adversarial robustness. Code: https://github.com/qizhangli/ST-AT.

Jian Zou, Tianyu Huang, Guanglei Yang et al.

Masked Autoencoders (MAE) play a pivotal role in learning potent representations, delivering outstanding results across various 3D perception tasks essential for autonomous driving. In real-world driving scenarios, it's commonplace to deploy multiple sensors for comprehensive environment perception. Despite integrating multi-modal features from these sensors can produce rich and powerful features, there is a noticeable challenge in MAE methods addressing this integration due to the substantial disparity between the different modalities. This research delves into multi-modal Masked Autoencoders tailored for a unified representation space in autonomous driving, aiming to pioneer a more efficient fusion of two distinct modalities. To intricately marry the semantics inherent in images with the geometric intricacies of LiDAR point clouds, we propose UniM$^2$AE. This model stands as a potent yet straightforward, multi-modal self-supervised pre-training framework, mainly consisting of two designs. First, it projects the features from both modalities into a cohesive 3D volume space to intricately marry the bird's eye view (BEV) with the height dimension. The extension allows for a precise representation of objects and reduces information loss when aligning multi-modal features. Second, the Multi-modal 3D Interactive Module (MMIM) is invoked to facilitate the efficient inter-modal interaction during the interaction process. Extensive experiments conducted on the nuScenes Dataset attest to the efficacy of UniM$^2$AE, indicating enhancements in 3D object detection and BEV map segmentation by 1.2\% NDS and 6.5\% mIoU, respectively. The code is available at https://github.com/hollow-503/UniM2AE.

Xin Zhong, Zhaoyi Yan, Jing Qin et al.

In recent years, crowd counting has become an important issue in computer vision. In most methods, the density maps are generated by convolving with a Gaussian kernel from the ground-truth dot maps which are marked around the center of human heads. Due to the fixed geometric structures in CNNs and indistinct head-scale information, the head features are obtained incompletely. Deformable convolution is proposed to exploit the scale-adaptive capabilities for CNN features in the heads. By learning the coordinate offsets of the sampling points, it is tractable to improve the ability to adjust the receptive field. However, the heads are not uniformly covered by the sampling points in the deformable convolution, resulting in loss of head information. To handle the non-uniformed sampling, an improved Normed-Deformable Convolution (\textit{i.e.,}NDConv) implemented by Normed-Deformable loss (\textit{i.e.,}NDloss) is proposed in this paper. The offsets of the sampling points which are constrained by NDloss tend to be more even. Then, the features in the heads are obtained more completely, leading to better performance. Especially, the proposed NDConv is a light-weight module which shares similar computation burden with Deformable Convolution. In the extensive experiments, our method outperforms state-of-the-art methods on ShanghaiTech A, ShanghaiTech B, UCF\_QNRF, and UCF\_CC\_50 dataset, achieving 61.4, 7.8, 91.2, and 167.2 MAE, respectively. The code is available at https://github.com/bingshuangzhuzi/NDConv

Yiwen Guo, Qizhang Li, Wangmeng Zuo et al.

This paper substantially extends our work published at ECCV, in which an intermediate-level attack was proposed to improve the transferability of some baseline adversarial examples. Specifically, we advocate a framework in which a direct linear mapping from the intermediate-level discrepancies (between adversarial features and benign features) to prediction loss of the adversarial example is established. By delving deep into the core components of such a framework, we show that 1) a variety of linear regression models can all be considered in order to establish the mapping, 2) the magnitude of the finally obtained intermediate-level adversarial discrepancy is correlated with the transferability, 3) further boost of the performance can be achieved by performing multiple runs of the baseline attack with random initialization. In addition, by leveraging these findings, we achieve new state-of-the-arts on transfer-based $\ell_\infty$ and $\ell_2$ attacks. Our code is publicly available at https://github.com/qizhangli/ila-plus-plus-lr.

Haolin Wang, Jiawei Zhang, Ming Liu et al.

Image retouching, aiming to regenerate the visually pleasing renditions of given images, is a subjective task where the users are with different aesthetic sensations. Most existing methods deploy a deterministic model to learn the retouching style from a specific expert, making it less flexible to meet diverse subjective preferences. Besides, the intrinsic diversity of an expert due to the targeted processing on different images is also deficiently described. To circumvent such issues, we propose to learn diverse image retouching with normalizing flow-based architectures. Unlike current flow-based methods which directly generate the output image, we argue that learning in a style domain could (i) disentangle the retouching styles from the image content, (ii) lead to a stable style presentation form, and (iii) avoid the spatial disharmony effects. For obtaining meaningful image tone style representations, a joint-training pipeline is delicately designed, which is composed of a style encoder, a conditional RetouchNet, and the image tone style normalizing flow (TSFlow) module. In particular, the style encoder predicts the target style representation of an input image, which serves as the conditional information in the RetouchNet for retouching, while the TSFlow maps the style representation vector into a Gaussian distribution in the forward pass. After training, the TSFlow can generate diverse image tone style vectors by sampling from the Gaussian distribution. Extensive experiments on MIT-Adobe FiveK and PPR10K datasets show that our proposed method performs favorably against state-of-the-art methods and is effective in generating diverse results to satisfy different human aesthetic preferences. Source code and pre-trained models are publicly available at https://github.com/SSRHeart/TSFlow.

Wei Shang, Dongwei Ren, Wanying Zhang et al.

Arbitrary-scale video super-resolution (AVSR) aims to enhance the resolution of video frames, potentially at various scaling factors, which presents several challenges regarding spatial detail reproduction, temporal consistency, and computational complexity. In this paper, we first describe a strong baseline for AVSR by putting together three variants of elementary building blocks: 1) a flow-guided recurrent unit that aggregates spatiotemporal information from previous frames, 2) a flow-refined cross-attention unit that selects spatiotemporal information from future frames, and 3) a hyper-upsampling unit that generates scaleaware and content-independent upsampling kernels. We then introduce ST-AVSR by equipping our baseline with a multi-scale structural and textural prior computed from the pre-trained VGG network. This prior has proven effective in discriminating structure and texture across different locations and scales, which is beneficial for AVSR. Comprehensive experiments show that ST-AVSR significantly improves super-resolution quality, generalization ability, and inference speed over the state-of-theart. The code is available at https://github.com/shangwei5/ST-AVSR.

Qinghe Wang, Lijie Liu, Miao Hua et al.

Image-based head swapping task aims to stitch a source head to another source body flawlessly. This seldom-studied task faces two major challenges: 1) Preserving the head and body from various sources while generating a seamless transition region. 2) No paired head swapping dataset and benchmark so far. In this paper, we propose a semantic-mixing diffusion model for head swapping (HS-Diffusion) which consists of a latent diffusion model (LDM) and a semantic layout generator. We blend the semantic layouts of source head and source body, and then inpaint the transition region by the semantic layout generator, achieving a coarse-grained head swapping. Semantic-mixing LDM can further implement a fine-grained head swapping with the inpainted layout as condition by a progressive fusion process, while preserving head and body with high-quality reconstruction. To this end, we propose a semantic calibration strategy for natural inpainting and a neck alignment for geometric realism. Importantly, we construct a new image-based head swapping benchmark and design two tailor-designed metrics (Mask-FID and Focal-FID). Extensive experiments demonstrate the superiority of our framework. The code will be available: https://github.com/qinghew/HS-Diffusion.

Wei Shang, Dongwei Ren, Yi Yang et al.

Video deblurring methods, aiming at recovering consecutive sharp frames from a given blurry video, usually assume that the input video suffers from consecutively blurry frames. However, in real-world scenarios captured by modern imaging devices, sharp frames often interspersed within the video, providing temporally nearest sharp features that can aid in the restoration of blurry frames. In this work, we propose a video deblurring method that leverages both neighboring frames and existing sharp frames using hybrid Transformers for feature aggregation. Specifically, we first train a blur-aware detector to distinguish between sharp and blurry frames. Then, a window-based local Transformer is employed for exploiting features from neighboring frames, where cross attention is beneficial for aggregating features from neighboring frames without explicit spatial alignment. To aggregate nearest sharp features from detected sharp frames, we utilize a global Transformer with multi-scale matching capability. Moreover, our method can easily be extended to event-driven video deblurring by incorporating an event fusion module into the global Transformer. Extensive experiments on benchmark datasets demonstrate that our proposed method outperforms state-of-the-art video deblurring methods as well as event-driven video deblurring methods in terms of quantitative metrics and visual quality. The source code and trained models are available at https://github.com/shangwei5/STGTN.

Zhicun Yin, Ming Liu, Xiaoming Li et al.

Due to their highly structured characteristics, faces are easier to recover than natural scenes for blind image super-resolution. Therefore, we can extract the degradation representation of an image from the low-quality and recovered face pairs. Using the degradation representation, realistic low-quality images can then be synthesized to fine-tune the super-resolution model for the real-world low-quality image. However, such a procedure is time-consuming and laborious, and the gaps between recovered faces and the ground-truths further increase the optimization uncertainty. To facilitate efficient model adaptation towards image-specific degradations, we propose a method dubbed MetaF2N, which leverages the contained Faces to fine-tune model parameters for adapting to the whole Natural image in a Meta-learning framework. The degradation extraction and low-quality image synthesis steps are thus circumvented in our MetaF2N, and it requires only one fine-tuning step to get decent performance. Considering the gaps between the recovered faces and ground-truths, we further deploy a MaskNet for adaptively predicting loss weights at different positions to reduce the impact of low-confidence areas. To evaluate our proposed MetaF2N, we have collected a real-world low-quality dataset with one or multiple faces in each image, and our MetaF2N achieves superior performance on both synthetic and real-world datasets. Source code, pre-trained models, and collected datasets are available at https://github.com/yinzhicun/MetaF2N.

Renlong Wu, Zhilu Zhang, Shuohao Zhang et al.

Burst super-resolution (BurstSR) aims at reconstructing a high-resolution (HR) image from a sequence of low-resolution (LR) and noisy images, which is conducive to enhancing the imaging effects of smartphones with limited sensors. The main challenge of BurstSR is to effectively combine the complementary information from input frames, while existing methods still struggle with it. In this paper, we suggest fusing cues frame-by-frame with an efficient and flexible recurrent network. In particular, we emphasize the role of the base-frame and utilize it as a key prompt to guide the knowledge acquisition from other frames in every recurrence. Moreover, we introduce an implicit weighting loss to improve the model's flexibility in facing input frames with variable numbers. Extensive experiments on both synthetic and real-world datasets demonstrate that our method achieves better results than state-of-the-art ones. Codes and pre-trained models are available at https://github.com/ZcsrenlongZ/RBSR.

Tianyi Zhu, Wei Shang, Dongwei Ren et al.

Animation line inbetweening is a crucial step in animation production aimed at enhancing animation fluidity by predicting intermediate line arts between two key frames. However, existing methods face challenges in effectively addressing sparse pixels and significant motion in line art key frames. In literature, Chamfer Distance (CD) is commonly adopted for evaluating inbetweening performance. Despite achieving favorable CD values, existing methods often generate interpolated frames with line disconnections, especially for scenarios involving large motion. Motivated by this observation, we propose a simple yet effective interpolation method for animation line inbetweening that adopts thin-plate spline-based transformation to estimate coarse motion more accurately by modeling the keypoint correspondence between two key frames, particularly for large motion scenarios. Building upon the coarse estimation, a motion refine module is employed to further enhance motion details before final frame interpolation using a simple UNet model. Furthermore, to more accurately assess the performance of animation line inbetweening, we refine the CD metric and introduce a novel metric termed Weighted Chamfer Distance, which demonstrates a higher consistency with visual perception quality. Additionally, we incorporate Earth Mover's Distance and conduct user study to provide a more comprehensive evaluation. Our method outperforms existing approaches by delivering high-quality interpolation results with enhanced fluidity. The code is available at \url{https://github.com/Tian-one/tps-inbetween}.

Mingwen Shao, Lingzhuang Meng, Yuanjian Qiao et al.

Since the training data of the target model is not available in the black-box substitute attack, most recent schemes utilize GANs to generate data for training the substitute model. However, these GANs-based schemes suffer from low training efficiency as the generator needs to be retrained for each target model during the substitute training process, as well as low generation quality. To overcome these limitations, we consider utilizing the diffusion model to generate data, and propose a novel data-free substitute attack scheme based on the Stable Diffusion (SD) to improve the efficiency and accuracy of substitute training. Despite the data generated by the SD exhibiting high quality, it presents a different distribution of domains and a large variation of positive and negative samples for the target model. For this problem, we propose Latent Code Augmentation (LCA) to facilitate SD in generating data that aligns with the data distribution of the target model. Specifically, we augment the latent codes of the inferred member data with LCA and use them as guidance for SD. With the guidance of LCA, the data generated by the SD not only meets the discriminative criteria of the target model but also exhibits high diversity. By utilizing this data, it is possible to train the substitute model that closely resembles the target model more efficiently. Extensive experiments demonstrate that our LCA achieves higher attack success rates and requires fewer query budgets compared to GANs-based schemes for different target models. Our codes are available at \url{https://github.com/LzhMeng/LCA}.

Ming Liu, Jianan Pan, Zifei Yan et al.

Recently, multiple synthetic and real-world datasets have been built to facilitate the training of deep single image reflection removal (SIRR) models. Meanwhile, diverse testing sets are also provided with different types of reflection and scenes. However, the non-negligible domain gaps between training and testing sets make it difficult to learn deep models generalizing well to testing images. The diversity of reflections and scenes further makes it a mission impossible to learn a single model being effective to all testing sets and real-world reflections. In this paper, we tackle these issues by learning SIRR models from a domain generalization perspective. Particularly, for each source set, a specific SIRR model is trained to serve as a domain expert of relevant reflection types. For a given reflection-contaminated image, we present a reflection type-aware weighting (RTAW) module to predict expert-wise weights. RTAW can then be incorporated with adaptive network combination (AdaNEC) for handling different reflection types and scenes, i.e., generalizing to unknown domains. Two representative AdaNEC methods, i.e., output fusion (OF) and network interpolation (NI), are provided by considering both adaptation levels and efficiency. For images from one source set, we train RTAW to only predict expert-wise weights of other domain experts for improving generalization ability, while the weights of all experts are predicted and employed during testing. An in-domain expert (IDE) loss is presented for training RTAW. Extensive experiments show the appealing performance gain of our AdaNEC on different state-of-the-art SIRR networks. Source code and pre-trained models will available at https://github.com/csmliu/AdaNEC.

Chunwei Tian, Menghua Zheng, Wangmeng Zuo et al.

Deep convolutional neural networks (CNNs) depend on feedforward and feedback ways to obtain good performance in image denoising. However, how to obtain effective structural information via CNNs to efficiently represent given noisy images is key for complex scenes. In this paper, we propose a cross Transformer denoising CNN (CTNet) with a serial block (SB), a parallel block (PB), and a residual block (RB) to obtain clean images for complex scenes. A SB uses an enhanced residual architecture to deeply search structural information for image denoising. To avoid loss of key information, PB uses three heterogeneous networks to implement multiple interactions of multi-level features to broadly search for extra information for improving the adaptability of an obtained denoiser for complex scenes. Also, to improve denoising performance, Transformer mechanisms are embedded into the SB and PB to extract complementary salient features for effectively removing noise in terms of pixel relations. Finally, a RB is applied to acquire clean images. Experiments illustrate that our CTNet is superior to some popular denoising methods in terms of real and synthetic image denoising. It is suitable to mobile digital devices, i.e., phones. Codes can be obtained at https://github.com/hellloxiaotian/CTNet.

Qizhang Li, Yiwen Guo, Xiaochen Yang et al.

The transferability of adversarial examples allows for the attack on unknown deep neural networks (DNNs), posing a serious threat to many applications and attracting great attention. In this paper, we improve the transferability of adversarial examples by incorporating the Bayesian formulation into both the model parameters and model input, enabling their joint diversification. We demonstrate that combination of Bayesian formulations for both the model input and model parameters yields significant improvements in transferability. By introducing advanced approximations of the posterior distribution over the model input, adversarial transferability achieves further enhancement, surpassing all state-of-the-arts when attacking without model fine-tuning. Additionally, we propose a principled approach to fine-tune model parameters within this Bayesian framework. Extensive experiments demonstrate that our method achieves a new state-of-the-art in transfer-based attacks, significantly improving the average success rate on ImageNet and CIFAR-10. Code at: https://github.com/qizhangli/MoreBayesian-jrnl.

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.

Yang Bai, Xinxing Xu, Yong Liu et al.

Composed image retrieval (CIR) is the task of retrieving specific images by using a query that involves both a reference image and a relative caption. Most existing CIR models adopt the late-fusion strategy to combine visual and language features. Besides, several approaches have also been suggested to generate a pseudo-word token from the reference image, which is further integrated into the relative caption for CIR. However, these pseudo-word-based prompting methods have limitations when target image encompasses complex changes on reference image, e.g., object removal and attribute modification. In this work, we demonstrate that learning an appropriate sentence-level prompt for the relative caption (SPRC) is sufficient for achieving effective composed image retrieval. Instead of relying on pseudo-word-based prompts, we propose to leverage pretrained V-L models, e.g., BLIP-2, to generate sentence-level prompts. By concatenating the learned sentence-level prompt with the relative caption, one can readily use existing text-based image retrieval models to enhance CIR performance. Furthermore, we introduce both image-text contrastive loss and text prompt alignment loss to enforce the learning of suitable sentence-level prompts. Experiments show that our proposed method performs favorably against the state-of-the-art CIR methods on the Fashion-IQ and CIRR datasets. The source code and pretrained model are publicly available at https://github.com/chunmeifeng/SPRC

Chun-Mei Feng, Kai Yu, Yong Liu et al.

Benefiting from prompt tuning, recent years have witnessed the promising performance of pre-trained vision-language models, e.g., CLIP, on versatile downstream tasks. In this paper, we focus on a particular setting of learning adaptive prompts on the fly for each test sample from an unseen new domain, which is known as test-time prompt tuning (TPT). Existing TPT methods typically rely on data augmentation and confidence selection. However, conventional data augmentation techniques, e.g., random resized crops, suffers from the lack of data diversity, while entropy-based confidence selection alone is not sufficient to guarantee prediction fidelity. To address these issues, we propose a novel TPT method, named DiffTPT, which leverages pre-trained diffusion models to generate diverse and informative new data. Specifically, we incorporate augmented data by both conventional method and pre-trained stable diffusion to exploit their respective merits, improving the models ability to adapt to unknown new test data. Moreover, to ensure the prediction fidelity of generated data, we introduce a cosine similarity-based filtration technique to select the generated data with higher similarity to the single test sample. Our experiments on test datasets with distribution shifts and unseen categories demonstrate that DiffTPT improves the zero-shot accuracy by an average of 5.13\% compared to the state-of-the-art TPT method. Our code and models will be publicly released.

Wei Shang, Dongwei Ren, Wanying Zhang et al.

Modern consumer cameras commonly employ the rolling shutter (RS) imaging mechanism, via which images are captured by scanning scenes row-by-row, resulting in RS distortion for dynamic scenes. To correct RS distortion, existing methods adopt a fully supervised learning manner that requires high framerate global shutter (GS) images as ground-truth for supervision. In this paper, we propose an enhanced Self-supervised learning framework for Dual reversed RS distortion Correction (SelfDRSC++). Firstly, we introduce a lightweight DRSC network that incorporates a bidirectional correlation matching block to refine the joint optimization of optical flows and corrected RS features, thereby improving correction performance while reducing network parameters. Subsequently, to effectively train the DRSC network, we propose a self-supervised learning strategy that ensures cycle consistency between input and reconstructed dual reversed RS images. The RS reconstruction in SelfDRSC++ can be interestingly formulated as a specialized instance of video frame interpolation, where each row in reconstructed RS images is interpolated from predicted GS images by utilizing RS distortion time maps. By achieving superior performance while simplifying the training process, SelfDRSC++ enables feasible one-stage self-supervised training. Additionally, besides start and end RS scanning time, SelfDRSC++ allows supervision of GS images at arbitrary intermediate scanning times, thus enabling the learned DRSC network to generate high framerate GS videos. The code and trained models are available at \url{https://github.com/shangwei5/SelfDRSC_plusplus}.

Bowen Dong, Pan Zhou, Shuicheng Yan et al.

For long-tailed classification, most works often pretrain a big model on a large-scale dataset, and then fine-tune the whole model for adapting to long-tailed data. Though promising, fine-tuning the whole pretrained model tends to suffer from high cost in computation and deployment of different models for different tasks, as well as weakened generalization ability for overfitting to certain features of long-tailed data. To alleviate these issues, we propose an effective Long-tailed Prompt Tuning method for long-tailed classification. LPT introduces several trainable prompts into a frozen pretrained model to adapt it to long-tailed data. For better effectiveness, we divide prompts into two groups: 1) a shared prompt for the whole long-tailed dataset to learn general features and to adapt a pretrained model into target domain; and 2) group-specific prompts to gather group-specific features for the samples which have similar features and also to empower the pretrained model with discrimination ability. Then we design a two-phase training paradigm to learn these prompts. In phase 1, we train the shared prompt via supervised prompt tuning to adapt a pretrained model to the desired long-tailed domain. In phase 2, we use the learnt shared prompt as query to select a small best matched set for a group of similar samples from the group-specific prompt set to dig the common features of these similar samples, then optimize these prompts with dual sampling strategy and asymmetric GCL loss. By only fine-tuning a few prompts while fixing the pretrained model, LPT can reduce training and deployment cost by storing a few prompts, and enjoys a strong generalization ability of the pretrained model. Experiments show that on various long-tailed benchmarks, with only ~1.1% extra parameters, LPT achieves comparable performance than previous whole model fine-tuning methods, and is more robust to domain-shift.

Yuanze Li, Haolin Wang, Shihao Yuan et al.

Due to the training configuration, traditional industrial anomaly detection (IAD) methods have to train a specific model for each deployment scenario, which is insufficient to meet the requirements of modern design and manufacturing. On the contrary, large multimodal models~(LMMs) have shown eminent generalization ability on various vision tasks, and their perception and comprehension capabilities imply the potential of applying LMMs on IAD tasks. However, we observe that even though the LMMs have abundant knowledge about industrial anomaly detection in the textual domain, the LMMs are unable to leverage the knowledge due to the modality gap between textual and visual domains. To stimulate the relevant knowledge in LMMs and adapt the LMMs towards anomaly detection tasks, we introduce existing IAD methods as vision experts and present a novel large multimodal model applying vision experts for industrial anomaly detection~(abbreviated to {Myriad}). Specifically, we utilize the anomaly map generated by the vision experts as guidance for LMMs, such that the vision model is guided to pay more attention to anomalous regions. Then, the visual features are modulated via an adapter to fit the anomaly detection tasks, which are fed into the language model together with the vision expert guidance and human instructions to generate the final outputs. Extensive experiments are applied on MVTec-AD, VisA, and PCB Bank benchmarks demonstrate that our proposed method not only performs favorably against state-of-the-art methods, but also inherits the flexibility and instruction-following ability of LMMs in the field of IAD. Source code and pre-trained models are publicly available at \url{https://github.com/tzjtatata/Myriad}.

Chun-Mei Feng, Bangjun Li, Xinxing Xu et al.

Federated Magnetic Resonance Imaging (MRI) reconstruction enables multiple hospitals to collaborate distributedly without aggregating local data, thereby protecting patient privacy. However, the data heterogeneity caused by different MRI protocols, insufficient local training data, and limited communication bandwidth inevitably impair global model convergence and updating. In this paper, we propose a new algorithm, FedPR, to learn federated visual prompts in the null space of global prompt for MRI reconstruction. FedPR is a new federated paradigm that adopts a powerful pre-trained model while only learning and communicating the prompts with few learnable parameters, thereby significantly reducing communication costs and achieving competitive performance on limited local data. Moreover, to deal with catastrophic forgetting caused by data heterogeneity, FedPR also updates efficient federated visual prompts that project the local prompts into an approximate null space of the global prompt, thereby suppressing the interference of gradients on the server performance. Extensive experiments on federated MRI show that FedPR significantly outperforms state-of-the-art FL algorithms with <6% of communication costs when given the limited amount of local training data.

Bowen Dong, Pan Zhou, Shuicheng Yan et al.

The few-shot learning ability of vision transformers (ViTs) is rarely investigated though heavily desired. In this work, we empirically find that with the same few-shot learning frameworks, \eg~Meta-Baseline, replacing the widely used CNN feature extractor with a ViT model often severely impairs few-shot classification performance. Moreover, our empirical study shows that in the absence of inductive bias, ViTs often learn the low-qualified token dependencies under few-shot learning regime where only a few labeled training data are available, which largely contributes to the above performance degradation. To alleviate this issue, for the first time, we propose a simple yet effective few-shot training framework for ViTs, namely Self-promoted sUpervisioN (SUN). Specifically, besides the conventional global supervision for global semantic learning SUN further pretrains the ViT on the few-shot learning dataset and then uses it to generate individual location-specific supervision for guiding each patch token. This location-specific supervision tells the ViT which patch tokens are similar or dissimilar and thus accelerates token dependency learning. Moreover, it models the local semantics in each patch token to improve the object grounding and recognition capability which helps learn generalizable patterns. To improve the quality of location-specific supervision, we further propose two techniques:~1) background patch filtration to filtrate background patches out and assign them into an extra background class; and 2) spatial-consistent augmentation to introduce sufficient diversity for data augmentation while keeping the accuracy of the generated local supervisions. Experimental results show that SUN using ViTs significantly surpasses other few-shot learning frameworks with ViTs and is the first one that achieves higher performance than those CNN state-of-the-arts.

Xiaoming Li, Shiguang Zhang, Shangchen Zhou et al.

To improve the performance of blind face restoration, recent works mainly treat the two aspects, i.e., generic and specific restoration, separately. In particular, generic restoration attempts to restore the results through general facial structure prior, while on the one hand, cannot generalize to real-world degraded observations due to the limited capability of direct CNNs' mappings in learning blind restoration, and on the other hand, fails to exploit the identity-specific details. On the contrary, specific restoration aims to incorporate the identity features from the reference of the same identity, in which the requirement of proper reference severely limits the application scenarios. Generally, it is a challenging and intractable task to improve the photo-realistic performance of blind restoration and adaptively handle the generic and specific restoration scenarios with a single unified model. Instead of implicitly learning the mapping from a low-quality image to its high-quality counterpart, this paper suggests a DMDNet by explicitly memorizing the generic and specific features through dual dictionaries. First, the generic dictionary learns the general facial priors from high-quality images of any identity, while the specific dictionary stores the identity-belonging features for each person individually. Second, to handle the degraded input with or without specific reference, dictionary transform module is suggested to read the relevant details from the dual dictionaries which are subsequently fused into the input features. Finally, multi-scale dictionaries are leveraged to benefit the coarse-to-fine restoration. Moreover, a new high-quality dataset, termed CelebRef-HQ, is constructed to promote the exploration of specific face restoration in the high-resolution space.

Xiaoming Li, Wangmeng Zuo, Chen Change Loy

Blind text image super-resolution (SR) is challenging as one needs to cope with diverse font styles and unknown degradation. To address the problem, existing methods perform character recognition in parallel to regularize the SR task, either through a loss constraint or intermediate feature condition. Nonetheless, the high-level prior could still fail when encountering severe degradation. The problem is further compounded given characters of complex structures, e.g., Chinese characters that combine multiple pictographic or ideographic symbols into a single character. In this work, we present a novel prior that focuses more on the character structure. In particular, we learn to encapsulate rich and diverse structures in a StyleGAN and exploit such generative structure priors for restoration. To restrict the generative space of StyleGAN so that it obeys the structure of characters yet remains flexible in handling different font styles, we store the discrete features for each character in a codebook. The code subsequently drives the StyleGAN to generate high-resolution structural details to aid text SR. Compared to priors based on character recognition, the proposed structure prior exerts stronger character-specific guidance to restore faithful and precise strokes of a designated character. Extensive experiments on synthetic and real datasets demonstrate the compelling performance of the proposed generative structure prior in facilitating robust text SR.

Yue Cao, Zhaolin Wan, Dongwei Ren et al.

Full-reference (FR) image quality assessment (IQA) evaluates the visual quality of a distorted image by measuring its perceptual difference with pristine-quality reference, and has been widely used in low-level vision tasks. Pairwise labeled data with mean opinion score (MOS) are required in training FR-IQA model, but is time-consuming and cumbersome to collect. In contrast, unlabeled data can be easily collected from an image degradation or restoration process, making it encouraging to exploit unlabeled training data to boost FR-IQA performance. Moreover, due to the distribution inconsistency between labeled and unlabeled data, outliers may occur in unlabeled data, further increasing the training difficulty. In this paper, we suggest to incorporate semi-supervised and positive-unlabeled (PU) learning for exploiting unlabeled data while mitigating the adverse effect of outliers. Particularly, by treating all labeled data as positive samples, PU learning is leveraged to identify negative samples (i.e., outliers) from unlabeled data. Semi-supervised learning (SSL) is further deployed to exploit positive unlabeled data by dynamically generating pseudo-MOS. We adopt a dual-branch network including reference and distortion branches. Furthermore, spatial attention is introduced in the reference branch to concentrate more on the informative regions, and sliced Wasserstein distance is used for robust difference map computation to address the misalignment issues caused by images recovered by GAN models. Extensive experiments show that our method performs favorably against state-of-the-arts on the benchmark datasets PIPAL, KADID-10k, TID2013, LIVE and CSIQ.

Yupeng Shi, Xiao Liu, Yuxiang Wei et al.

Semantic image synthesis is a challenging task with many practical applications. Albeit remarkable progress has been made in semantic image synthesis with spatially-adaptive normalization and existing methods normalize the feature activations under the coarse-level guidance (e.g., semantic class). However, different parts of a semantic object (e.g., wheel and window of car) are quite different in structures and textures, making blurry synthesis results usually inevitable due to the missing of fine-grained guidance. In this paper, we propose a novel normalization module, termed as REtrieval-based Spatially AdaptIve normaLization (RESAIL), for introducing pixel level fine-grained guidance to the normalization architecture. Specifically, we first present a retrieval paradigm by finding a content patch of the same semantic class from training set with the most similar shape to each test semantic mask. Then, RESAIL is presented to use the retrieved patch for guiding the feature normalization of corresponding region, and can provide pixel level fine-grained guidance, thereby greatly mitigating blurry synthesis results. Moreover, distorted ground-truth images are also utilized as alternatives of retrieval-based guidance for feature normalization, further benefiting model training and improving visual quality of generated images. Experiments on several challenging datasets show that our RESAIL performs favorably against state-of-the-arts in terms of quantitative metrics, visual quality, and subjective evaluation. The source code and pre-trained models will be publicly available.

Mingxiang Liao, Hannan Lu, Xinyu Zhang et al.

Comprehensive and constructive evaluation protocols play an important role in the development of sophisticated text-to-video (T2V) generation models. Existing evaluation protocols primarily focus on temporal consistency and content continuity, yet largely ignore the dynamics of video content. Dynamics are an essential dimension for measuring the visual vividness and the honesty of video content to text prompts. In this study, we propose an effective evaluation protocol, termed DEVIL, which centers on the dynamics dimension to evaluate T2V models. For this purpose, we establish a new benchmark comprising text prompts that fully reflect multiple dynamics grades, and define a set of dynamics scores corresponding to various temporal granularities to comprehensively evaluate the dynamics of each generated video. Based on the new benchmark and the dynamics scores, we assess T2V models with the design of three metrics: dynamics range, dynamics controllability, and dynamics-based quality. Experiments show that DEVIL achieves a Pearson correlation exceeding 90% with human ratings, demonstrating its potential to advance T2V generation models. Code is available at https://github.com/MingXiangL/DEVIL.

Zhilu Zhang, Haoyu Wang, Shuai Liu et al.

Merging multi-exposure images is a common approach for obtaining high dynamic range (HDR) images, with the primary challenge being the avoidance of ghosting artifacts in dynamic scenes. Recent methods have proposed using deep neural networks for deghosting. However, the methods typically rely on sufficient data with HDR ground-truths, which are difficult and costly to collect. In this work, to eliminate the need for labeled data, we propose SelfHDR, a self-supervised HDR reconstruction method that only requires dynamic multi-exposure images during training. Specifically, SelfHDR learns a reconstruction network under the supervision of two complementary components, which can be constructed from multi-exposure images and focus on HDR color as well as structure, respectively. The color component is estimated from aligned multi-exposure images, while the structure one is generated through a structure-focused network that is supervised by the color component and an input reference (\eg, medium-exposure) image. During testing, the learned reconstruction network is directly deployed to predict an HDR image. Experiments on real-world images demonstrate our SelfHDR achieves superior results against the state-of-the-art self-supervised methods, and comparable performance to supervised ones. Codes are available at https://github.com/cszhilu1998/SelfHDR