Mingjie Wang

Mingjie Wang, Yande Li, Jun Zhou et al.

The class-agnostic counting (CAC) problem has caught increasing attention recently due to its wide societal applications and arduous challenges. To count objects of different categories, existing approaches rely on user-provided exemplars, which is hard-to-obtain and limits their generality. In this paper, we aim to empower the framework to recognize adaptive exemplars within the whole images. A zero-shot Generalized Counting Network (GCNet) is developed, which uses a pseudo-Siamese structure to automatically and effectively learn pseudo exemplar clues from inherent repetition patterns. In addition, a weakly-supervised scheme is presented to reduce the burden of laborious density maps required by all contemporary CAC models, allowing GCNet to be trained using count-level supervisory signals in an end-to-end manner. Without providing any spatial location hints, GCNet is capable of adaptively capturing them through a carefully-designed self-similarity learning strategy. Extensive experiments and ablation studies on the prevailing benchmark FSC147 for zero-shot CAC demonstrate the superiority of our GCNet. It performs on par with existing exemplar-dependent methods and shows stunning cross-dataset generality on crowd-specific datasets, e.g., ShanghaiTech Part A, Part B and UCF_QNRF.

Mingjie Wang, Jianxiong Guo, Sirui Li et al.

Deep neural networks have significantly advanced person re-identification (ReID) applications in the realm of the industrial internet, yet they remain vulnerable. Thus, it is crucial to study the robustness of ReID systems, as there are risks of adversaries using these vulnerabilities to compromise industrial surveillance systems. Current adversarial methods focus on generating attack samples using misclassification feedback from victim models (VMs), neglecting VM's cognitive processes. We seek to address this by producing authentic ReID attack instances through VM cognition decryption. This approach boasts advantages like better transferability to open-set ReID tests, easier VM misdirection, and enhanced creation of realistic and undetectable assault images. However, the task of deciphering the cognitive mechanism in VM is widely considered to be a formidable challenge. In this paper, we propose a novel inconspicuous and controllable ReID attack baseline, LCYE (Look Closer to Your Enemy), to generate adversarial query images. Specifically, LCYE first distills VM's knowledge via teacher-student memory mimicking the proxy task. This knowledge prior serves as an unambiguous cryptographic token, encapsulating elements deemed indispensable and plausible by the VM, with the intent of facilitating precise adversarial misdirection. Further, benefiting from the multiple opposing task framework of LCYE, we investigate the interpretability and generalization of ReID models from the view of the adversarial attack, including cross-domain adaption, cross-model consensus, and online learning process. Extensive experiments on four ReID benchmarks show that our method outperforms other state-of-the-art attackers with a large margin in white-box, black-box, and target attacks. The source code can be found at https://github.com/MingjieWang0606/LCYE-attack_reid.

Mingjie Wang, Jun Zhou, Hao Cai et al.

Existing state-of-the-art crowd counting algorithms rely excessively on location-level annotations, which are burdensome to acquire. When only count-level (weak) supervisory signals are available, it is arduous and error-prone to regress total counts due to the lack of explicit spatial constraints. To address this issue, a novel and efficient counter (referred to as CrowdMLP) is presented, which probes into modelling global dependencies of embeddings and regressing total counts by devising a multi-granularity MLP regressor. In specific, a locally-focused pre-trained frontend is cascaded to extract crude feature maps with intrinsic spatial cues, which prevent the model from collapsing into trivial outcomes. The crude embeddings, along with raw crowd scenes, are tokenized at different granularity levels. The multi-granularity MLP then proceeds to mix tokens at the dimensions of cardinality, channel, and spatial for mining global information. An effective proxy task, namely Split-Counting, is also proposed to evade the barrier of limited samples and the shortage of spatial hints in a self-supervised manner. Extensive experiments demonstrate that CrowdMLP significantly outperforms existing weakly-supervised counting algorithms and performs on par with state-of-the-art location-level supervised approaches.

Yande Li, Mingjie Wang, Minglun Gong et al.

The ever-increasing demands for intuitive interactions in Virtual Reality has triggered a boom in the realm of Facial Expression Recognition (FER). To address the limitations in existing approaches (e.g., narrow receptive fields and homogenous supervisory signals) and further cement the capacity of FER tools, a novel multifarious supervision-steering Transformer for FER in the wild is proposed in this paper. Referred as FER-former, our approach features multi-granularity embedding integration, hybrid self-attention scheme, and heterogeneous domain-steering supervision. In specific, to dig deep into the merits of the combination of features provided by prevailing CNNs and Transformers, a hybrid stem is designed to cascade two types of learning paradigms simultaneously. Wherein, a FER-specific transformer mechanism is devised to characterize conventional hard one-hot label-focusing and CLIP-based text-oriented tokens in parallel for final classification. To ease the issue of annotation ambiguity, a heterogeneous domains-steering supervision module is proposed to make image features also have text-space semantic correlations by supervising the similarity between image features and text features. On top of the collaboration of multifarious token heads, diverse global receptive fields with multi-modal semantic cues are captured, thereby delivering superb learning capability. Extensive experiments on popular benchmarks demonstrate the superiority of the proposed FER-former over the existing state-of-the-arts.

Wei Song, Jun Zhou, Mingjie Wang et al.

This paper focuses on the recently popular task of point cloud completion guided by multimodal information. Although existing methods have achieved excellent performance by fusing auxiliary images, there are still some deficiencies, including the poor generalization ability of the model and insufficient fine-grained semantic information for extracted features. In this work, we propose a novel multimodal fusion network for point cloud completion, which can simultaneously fuse visual and textual information to predict the semantic and geometric characteristics of incomplete shapes effectively. Specifically, to overcome the lack of prior information caused by the small-scale dataset, we employ a pre-trained vision-language model that is trained with a large amount of image-text pairs. Therefore, the textual and visual encoders of this large-scale model have stronger generalization ability. Then, we propose a multi-stage feature fusion strategy to fuse the textual and visual features into the backbone network progressively. Meanwhile, to further explore the effectiveness of fine-grained text descriptions for point cloud completion, we also build a text corpus with fine-grained descriptions, which can provide richer geometric details for 3D shapes. The rich text descriptions can be used for training and evaluating our network. Extensive quantitative and qualitative experiments demonstrate the superior performance of our method compared to state-of-the-art point cloud completion networks.

Mingjie Wang, Jianxiong Guo, Weijia Jia

Federated Learning (FL) is a decentralized learning method used to train machine learning algorithms. In FL, a global model iteratively collects the parameters of local models without accessing their local data. However, a significant challenge in FL is handling the heterogeneity of local data distribution, which often results in a drifted global model that is difficult to converge. To address this issue, current methods employ different strategies such as knowledge distillation, weighted model aggregation, and multi-task learning. These approaches are referred to as asynchronous FL, as they align user models either locally or post-hoc, where model drift has already occurred or has been underestimated. In this paper, we propose an active and synchronous correlation approach to address the challenge of user heterogeneity in FL. Specifically, our approach aims to approximate FL as standard deep learning by actively and synchronously scheduling user learning pace in each round with a dynamic multi-phase curriculum. A global curriculum is formed by an auto-regressive auto-encoder that integrates all user curricula on the server. This global curriculum is then divided into multiple phases and broadcast to users to measure and align the domain-agnostic learning pace. Empirical studies demonstrate that our approach outperforms existing asynchronous approaches in terms of generalization performance, even in the presence of severe user heterogeneity.

Xinyu Chen, Yuyi Qian, Jiang Lin et al.

Video object insertion requires ensuring spatio-temporal coherence and interactive realism, extending far beyond simple content placement. However, current approaches are often hindered by a reliance on explicit motion engineering or resource-intensive retraining, restricting their flexibility and generalization. To bridge this gap, we present \textit{SimInsert}, a training-free paradigm that efficiently decouples the task into intuitive single-frame editing and semantic motion description. By harnessing the robust generative priors of image-to-video diffusion models, SimInsert propagates edits temporally, strictly preserving background invariance while enabling plausible, text-driven interactions between the inserted object and the dynamic environment. Our approach hinges on non-invasive guidance mechanisms that enforce structural consistency, facilitate seamless boundary fusion, and counteract the fidelity drift that typically accumulates during the denoising trajectory. Extensive quantitative experiments validate our efficacy: SimInsert surpasses state-of-the-art methods with an 18.8\% gain in PSNR, 20.1\% in SSIM, and a 44.1\% decrease in LPIPS, offering a streamlined solution for high-fidelity video editing.

Dongqi Fan, Tao Chen, Mingjie Wang et al.

Current Pose-Guided Person Image Synthesis (PGPIS) methods depend heavily on large amounts of labeled triplet data to train the generator in a supervised manner. However, they often falter when applied to in-the-wild samples, primarily due to the distribution gap between the training datasets and real-world test samples. While some researchers aim to enhance model generalizability through sophisticated training procedures, advanced architectures, or by creating more diverse datasets, we adopt the test-time fine-tuning paradigm to customize a pre-trained Text2Image (T2I) model. However, naively applying test-time tuning results in inconsistencies in facial identities and appearance attributes. To address this, we introduce a Visual Consistency Module (VCM), which enhances appearance consistency by combining the face, text, and image embedding. Our approach, named OnePoseTrans, requires only a single source image to generate high-quality pose transfer results, offering greater stability than state-of-the-art data-driven methods. For each test case, OnePoseTrans customizes a model in around 48 seconds with an NVIDIA V100 GPU.

Zhiqiu Zhang, Dongqi Fan, Mingjie Wang et al.

The goal of image harmonization is to adjust the foreground in a composite image to achieve visual consistency with the background. Recently, latent diffusion model (LDM) are applied for harmonization, achieving remarkable results. However, LDM-based harmonization faces challenges in detail preservation and limited harmonization ability. Additionally, current synthetic datasets rely on color transfer, which lacks local variations and fails to capture complex real-world lighting conditions. To enhance harmonization capabilities, we propose the Region-to-Region transformation. By injecting information from appropriate regions into the foreground, this approach preserves original details while achieving image harmonization or, conversely, generating new composite data. From this perspective, We propose a novel model R2R. Specifically, we design Clear-VAE to preserve high-frequency details in the foreground using Adaptive Filter while eliminating disharmonious elements. To further enhance harmonization, we introduce the Harmony Controller with Mask-aware Adaptive Channel Attention (MACA), which dynamically adjusts the foreground based on the channel importance of both foreground and background regions. To address the limitation of existing datasets, we propose Random Poisson Blending, which transfers color and lighting information from a suitable region to the foreground, thereby generating more diverse and challenging synthetic images. Using this method, we construct a new synthetic dataset, RPHarmony. Experiments demonstrate the superiority of our method over other methods in both quantitative metrics and visual harmony. Moreover, our dataset helps the model generate more realistic images in real examples. Our code, dataset, and model weights have all been released for open access.

Tianyidan Xie, Zhentao Huang, Mingjie Wang et al.

Existing image-to-video generation methods often produce physically implausible motions and lack precise control over object dynamics. While prior approaches have incorporated physics simulators, they remain confined to 2D planar motions and fail to capture depth-aware spatial interactions. We introduce PhysLayer, a novel framework enabling language-guided, depth-aware layered animation of static images. PhysLayer consists of three key components: First, a language-guided scene understanding module that utilizes vision foundation models to decompose scenes into depth-based layers by analyzing object composition, material properties, and physical parameters. Second, a depth-aware layered physics simulation that extends 2D rigid-body dynamics with depth motion and perspective-consistent scaling, enabling more realistic object interactions without requiring full 3D reconstruction. Third, a physics-guided video synthesis module that integrates simulated trajectories with scene-aware relighting for temporally coherent results. Experimental results demonstrate improvements in CLIP-Similarity (+2.2\%), FID score (+9.3\%), and Motion-FID (+3\%), with human evaluation showing enhanced physical plausibility (+24\%) and text-video alignment (+35\%). Our approach provides a practical balance between physical realism and computational efficiency for controllable image animation.

Tianyidan Xie, Zhentao Huang, Mingjie Wang et al.

Automated movie creation requires coordinating multiple characters, modalities, and narrative elements across extended sequences -- a challenge that existing end-to-end approaches struggle to address effectively. We present \textbf{CineAGI}, a hierarchical movie generation framework that decomposes this complex task through specialized multi-agent orchestration. Our framework employs three key innovations: (1) a multi-agent narrative synthesis module where specialized LLM agents collaboratively generate comprehensive cinematic blueprints with character profiles, scene descriptions, and cross-modal specifications; (2) a decoupled character-centric pipeline that maintains identity consistency through instance-level tracking and integration while enabling flexible multi-character composition; and (3) a hierarchical audio-visual synchronization mechanism ensuring frame-level alignment of dialogue, expressions, and music. Extensive experiments demonstrate that CineAGI achieves 40\% improvement in overall consistency, 4.4\% gain in subject consistency, 5.4\% enhancement in aesthetic quality, and 28.7\% higher character consistency compared to baselines. Our work establishes a principled foundation for automated multi-scene video generation that preserves narrative coherence and character authenticity.

Jun Zhou, Dinghao Li, Nannan Li et al.

Recent advancements in multi-view 3D reconstruction and novel-view synthesis, particularly through Neural Radiance Fields (NeRF) and 3D Gaussian Splatting (3DGS), have greatly enhanced the fidelity and efficiency of 3D content creation. However, inpainting 3D scenes remains a challenging task due to the inherent irregularity of 3D structures and the critical need for maintaining multi-view consistency. In this work, we propose a novel 3D Gaussian inpainting framework that reconstructs complete 3D scenes by leveraging sparse inpainted views. Our framework incorporates an automatic Mask Refinement Process and region-wise Uncertainty-guided Optimization. Specifically, we refine the inpainting mask using a series of operations, including Gaussian scene filtering and back-projection, enabling more accurate localization of occluded regions and realistic boundary restoration. Furthermore, our Uncertainty-guided Fine-grained Optimization strategy, which estimates the importance of each region across multi-view images during training, alleviates multi-view inconsistencies and enhances the fidelity of fine details in the inpainted results. Comprehensive experiments conducted on diverse datasets demonstrate that our approach outperforms existing state-of-the-art methods in both visual quality and view consistency.

Jun Zhou, Yaoshun Li, Hongchen Tan et al.

In recent years, point cloud normal estimation, as a classical and foundational algorithm, has garnered extensive attention in the field of 3D geometric processing. Despite the remarkable performance achieved by current Neural Network-based methods, their robustness is still influenced by the quality of training data and the models' performance. In this study, we designed a fundamental framework for normal estimation, enhancing existing model through the incorporation of global information and various constraint mechanisms. Additionally, we employed a confidence-based strategy to select the reasonable samples for fair and robust network training. The introduced sample confidence can be integrated into the loss function to balance the influence of different samples on model training. Finally, we utilized existing orientation methods to correct estimated non-oriented normals, achieving state-of-the-art performance in both oriented and non-oriented tasks. Extensive experimental results demonstrate that our method works well on the widely used benchmarks.

Jun Wang, Hao Ruan, Liangjian Wen et al.

Visual gaze estimation, with its wide-ranging application scenarios, has garnered increasing attention within the research community. Although existing approaches infer gaze solely from image signals, recent advances in visual-language collaboration have demonstrated that the integration of linguistic information can significantly enhance performance across various visual tasks. Leveraging the remarkable transferability of large-scale Contrastive Language-Image Pre-training (CLIP) models, we address the open and urgent question of how to effectively apply linguistic cues to gaze estimation. In this work, we propose GazeCLIP, a novel gaze estimation framework that deeply explores text-face collaboration. Specifically, we introduce a meticulously designed linguistic description generator to produce text signals enriched with coarse directional cues. Furthermore, we present a CLIP-based backbone adept at characterizing text-face pairs for gaze estimation, complemented by a fine-grained multimodal fusion module that models the intricate interrelationships between heterogeneous inputs. Extensive experiments on three challenging datasets demonstrate the superiority of GazeCLIP, which achieves state-of-the-art accuracy. Our findings underscore the potential of using visual-language collaboration to advance gaze estimation and open new avenues for future research in multimodal learning for visual tasks. The implementation code and the pre-trained model will be made publicly available.

Quanxiu Wang, Hui Huang, Mingjie Wang et al.

Over the past decade, a series of unflagging efforts have been dedicated to developing highly expressive and controllable text-to-speech (TTS) systems. In general, the holistic TTS comprises two interconnected components: the frontend module and the backend module. The frontend excels in capturing linguistic representations from the raw text input, while the backend module converts linguistic cues to speech. The research community has shown growing interest in the study of the frontend component, recognizing its pivotal role in text-to-speech systems, including Text Normalization (TN), Prosody Boundary Prediction (PBP), and Polyphone Disambiguation (PD). Nonetheless, the limitations posed by insufficient annotated textual data and the reliance on homogeneous text signals significantly undermine the effectiveness of its supervised learning. To evade this obstacle, a novel two-stage TTS frontend prediction pipeline, named TAP-FM, is proposed in this paper. Specifically, during the first learning phase, we present a Multi-scale Contrastive Text-audio Pre-training protocol (MC-TAP), which hammers at acquiring richer insights via multi-granularity contrastive pre-training in an unsupervised manner. Instead of mining homogeneous features in prior pre-training approaches, our framework demonstrates the ability to delve deep into both global and local text-audio semantic and acoustic representations. Furthermore, a parallelized TTS frontend model is delicately devised to execute TN, PD, and PBP prediction tasks, respectively in the second stage. Finally, extensive experiments illustrate the superiority of our proposed method, achieving state-of-the-art performance.

Mingjie Wang, Jun Zhou, Yong Dai et al.

Recently, Class-Agnostic Counting (CAC) problem has garnered increasing attention owing to its intriguing generality and superior efficiency compared to Category-Specific Counting (CSC). This paper proposes a novel ExpressCount to enhance zero-shot object counting by delving deeply into language-guided exemplar learning. Specifically, the ExpressCount is comprised of an innovative Language-oriented Exemplar Perceptron and a downstream visual Zero-shot Counting pipeline. Thereinto, the perceptron hammers at exploiting accurate exemplar cues from collaborative language-vision signals by inheriting rich semantic priors from the prevailing pre-trained Large Language Models (LLMs), whereas the counting pipeline excels in mining fine-grained features through dual-branch and cross-attention schemes, contributing to the high-quality similarity learning. Apart from building a bridge between the LLM in vogue and the visual counting tasks, expression-guided exemplar estimation significantly advances zero-shot learning capabilities for counting instances with arbitrary classes. Moreover, devising a FSC-147-Express with annotations of meticulous linguistic expressions pioneers a new venue for developing and validating language-based counting models. Extensive experiments demonstrate the state-of-the-art performance of our ExpressCount, even showcasing the accuracy on par with partial CSC models.

Jingning Xu, Benlai Tang, Mingjie Wang et al.

Recently, talking face generation has drawn ever-increasing attention from the research community in computer vision due to its arduous challenges and widespread application scenarios, e.g. movie animation and virtual anchor. Although persevering efforts have been undertaken to enhance the fidelity and lip-sync quality of generated talking face videos, there is still large room for further improvements of synthesis quality and efficiency. Actually, these attempts somewhat ignore the explorations of fine-granularity feature extraction/integration and the consistency between probability distributions of landmarks, thereby recurring the issues of local details blurring and degraded fidelity. To mitigate these dilemmas, in this paper, a novel CLIP-based Attention and Probability Map Guided Network (CPNet) is delicately designed for inferring high-fidelity talking face videos. Specifically, considering the demands of fine-grained feature recalibration, a clip-based attention condenser is exploited to transfer knowledge with rich semantic priors from the prevailing CLIP model. Moreover, to guarantee the consistency in probability space and suppress the landmark ambiguity, we creatively propose the density map of facial landmark as auxiliary supervisory signal to guide the landmark distribution learning of generated frame. Extensive experiments on the widely-used benchmark dataset demonstrate the superiority of our CPNet against state of the arts in terms of image and lip-sync quality. In addition, a cohort of studies are also conducted to ablate the impacts of the individual pivotal components.

Tianyi Xie, Liucheng Liao, Cheng Bi et al.

The task of few-shot visual dubbing focuses on synchronizing the lip movements with arbitrary speech input for any talking head video. Albeit moderate improvements in current approaches, they commonly require high-quality homologous data sources of videos and audios, thus causing the failure to leverage heterogeneous data sufficiently. In practice, it may be intractable to collect the perfect homologous data in some cases, for example, audio-corrupted or picture-blurry videos. To explore this kind of data and support high-fidelity few-shot visual dubbing, in this paper, we novelly propose a simple yet efficient two-stage framework with a higher flexibility of mining heterogeneous data. Specifically, our two-stage paradigm employs facial landmarks as intermediate prior of latent representations and disentangles the lip movements prediction from the core task of realistic talking head generation. By this means, our method makes it possible to independently utilize the training corpus for two-stage sub-networks using more available heterogeneous data easily acquired. Besides, thanks to the disentanglement, our framework allows a further fine-tuning for a given talking head, thereby leading to better speaker-identity preserving in the final synthesized results. Moreover, the proposed method can also transfer appearance features from others to the target speaker. Extensive experimental results demonstrate the superiority of our proposed method in generating highly realistic videos synchronized with the speech over the state-of-the-art.

Jingning Xu, Benlai Tang, Mingjie Wang et al.

Clothes style transfer for person video generation is a challenging task, due to drastic variations of intra-person appearance and video scenarios. To tackle this problem, most recent AdaIN-based architectures are proposed to extract clothes and scenario features for generation. However, these approaches suffer from being short of fine-grained details and are prone to distort the origin person. To further improve the generation performance, we propose a novel framework with disentangled multi-branch encoders and a shared decoder. Moreover, to pursue the strong video spatio-temporal consistency, an inner-frame discriminator is delicately designed with input being cross-frame difference. Besides, the proposed framework possesses the property of scenario adaptation. Extensive experiments on the TEDXPeople benchmark demonstrate the superiority of our method over state-of-the-art approaches in terms of image quality and video coherence.

Yiming Sun, Feng Chen, Zhiyu Chen et al.

Deep neural networks are found to be prone to adversarial examples which could deliberately fool the model to make mistakes. Recently, a few of works expand this task from 2D image to 3D point cloud by using global point cloud optimization. However, the perturbations of global point are not effective for misleading the victim model. First, not all points are important in optimization toward misleading. Abundant points account considerable distortion budget but contribute trivially to attack. Second, the multi-label optimization is suboptimal for adversarial attack, since it consumes extra energy in finding multi-label victim model collapse and causes instance transformation to be dissimilar to any particular instance. Third, the independent adversarial and perceptibility losses, caring misclassification and dissimilarity separately, treat the updating of each point equally without a focus. Therefore, once perceptibility loss approaches its budget threshold, all points would be stock in the surface of hypersphere and attack would be locked in local optimality. Therefore, we propose a local aggressive adversarial attacks (L3A) to solve above issues. Technically, we select a bunch of salient points, the high-score subset of point cloud according to gradient, to perturb. Then a flow of aggressive optimization strategies are developed to reinforce the unperceptive generation of adversarial examples toward misleading victim models. Extensive experiments on PointNet, PointNet++ and DGCNN demonstrate the state-of-the-art performance of our method against existing adversarial attack methods.

Jun Zhou, Wei Jin, Mingjie Wang et al.

With the burst development of neural networks in recent years, the task of normal estimation has once again become a concern. By introducing the neural networks to classic methods based on problem-specific knowledge, the adaptability of the normal estimation algorithm to noise and scale has been greatly improved. However, the compatibility between neural networks and the traditional methods has not been considered. Similar to the principle of Occam's razor, that is, the simpler is better. We observe that a more simplified process of surface fitting can significantly improve the accuracy of the normal estimation. In this paper, two simple-yet-effective strategies are proposed to address the compatibility between the neural networks and surface fitting process to improve normal estimation. Firstly, a dynamic top-k selection strategy is introduced to better focus on the most critical points of a given patch, and the points selected by our learning method tend to fit a surface by way of a simple tangent plane, which can dramatically improve the normal estimation results of patches with sharp corners or complex patterns. Then, we propose a point update strategy before local surface fitting, which smooths the sharp boundary of the patch to simplify the surface fitting process, significantly reducing the fitting distortion and improving the accuracy of the predicted point normal. The experiments analyze the effectiveness of our proposed strategies and demonstrate that our method achieves SOTA results with the advantage of higher estimation accuracy over most existed approaches.

Jun Zhou, Wei Jin, Mingjie Wang et al.

This paper presents an effective normal estimation method adopting multi-patch stitching for an unstructured point cloud. The majority of learning-based approaches encode a local patch around each point of a whole model and estimate the normals in a point-by-point manner. In contrast, we suggest a more efficient pipeline, in which we introduce a patch-level normal estimation architecture to process a series of overlapping patches. Additionally, a multi-normal selection method based on weights, dubbed as multi-patch stitching, integrates the normals from the overlapping patches. To reduce the adverse effects of sharp corners or noise in a patch, we introduce an adaptive local feature aggregation layer to focus on an anisotropic neighborhood. We then utilize a multi-branch planar experts module to break the mutual influence between underlying piecewise surfaces in a patch. At the stitching stage, we use the learned weights of multi-branch planar experts and distance weights between points to select the best normal from the overlapping parts. Furthermore, we put forward constructing a sparse matrix representation to reduce large-scale retrieval overheads for the loop iterations dramatically. Extensive experiments demonstrate that our method achieves SOTA results with the advantage of lower computational costs and higher robustness to noise over most of the existing approaches.

Mingjie Wang, Hao Cai, Xianfeng Han et al.

Crowd counting remains a challenging task because the presence of drastic scale variation, density inconsistency, and complex background can seriously degrade the counting accuracy. To battle the ingrained issue of accuracy degradation, we propose a novel and powerful network called Scale Tree Network (STNet) for accurate crowd counting. STNet consists of two key components: a Scale-Tree Diversity Enhancer and a Semi-supervised Multi-level Auxiliator. Specifically, the Diversity Enhancer is designed to enrich scale diversity, which alleviates limitations of existing methods caused by insufficient level of scales. A novel tree structure is adopted to hierarchically parse coarse-to-fine crowd regions. Furthermore, a simple yet effective Multi-level Auxiliator is presented to aid in exploiting generalisable shared characteristics at multiple levels, allowing more accurate pixel-wise background cognition. The overall STNet is trained in an end-to-end manner, without the needs for manually tuning loss weights between the main and the auxiliary tasks. Extensive experiments on four challenging crowd datasets demonstrate the superiority of the proposed method.

Mingjie Wang, Hao Cai, Jun Zhou et al.

Crowd counting is an important vision task, which faces challenges on continuous scale variation within a given scene and huge density shift both within and across images. These challenges are typically addressed using multi-column structures in existing methods. However, such an approach does not provide consistent improvement and transferability due to limited ability in capturing multi-scale features, sensitiveness to large density shift, and difficulty in training multi-branch models. To overcome these limitations, a Single-column Scale-invariant Network (ScSiNet) is presented in this paper, which extracts sophisticated scale-invariant features via the combination of interlayer multi-scale integration and a novel intralayer scale-invariant transformation (SiT). Furthermore, in order to enlarge the diversity of densities, a randomly integrated loss is presented for training our single-branch method. Extensive experiments on public datasets demonstrate that the proposed method consistently outperforms state-of-the-art approaches in counting accuracy and achieves remarkable transferability and scale-invariant property.

Mingjie Wang, Jun Zhou, Wendong Mao et al.

Recently, Convolution Neural Networks (CNNs) obtained huge success in numerous vision tasks. In particular, DenseNets have demonstrated that feature reuse via dense skip connections can effectively alleviate the difficulty of training very deep networks and that reusing features generated by the initial layers in all subsequent layers has strong impact on performance. To feed even richer information into the network, a novel adaptive Multi-scale Convolution Aggregation module is presented in this paper. Composed of layers for multi-scale convolutions, trainable cross-scale aggregation, maxout, and concatenation, this module is highly non-linear and can boost the accuracy of DenseNet while using much fewer parameters. In addition, due to high model complexity, the network with extremely dense feature reuse is prone to overfitting. To address this problem, a regularization method named Stochastic Feature Reuse is also presented. Through randomly dropping a set of feature maps to be reused for each mini-batch during the training phase, this regularization method reduces training costs and prevents co-adaptation. Experimental results on CIFAR-10, CIFAR-100 and SVHN benchmarks demonstrated the effectiveness of the proposed methods.

Wendong Mao, Mingjie Wang, Jun Zhou et al.

A robust solution for semi-dense stereo matching is presented. It utilizes two CNN models for computing stereo matching cost and performing confidence-based filtering, respectively. Compared to existing CNNs-based matching cost generation approaches, our method feeds additional global information into the network so that the learned model can better handle challenging cases, such as lighting changes and lack of textures. Through utilizing non-parametric transforms, our method is also more self-reliant than most existing semi-dense stereo approaches, which rely highly on the adjustment of parameters. The experimental results based on Middlebury Stereo dataset demonstrate that the proposed approach outperforms the state-of-the-art semi-dense stereo approaches.