Yuhang Lu

Yue Jiang, Xue Jiang, Lihua Zhang et al.

Multimodal large language models (MLLMs) demonstrate remarkable visual understanding, yet their reliability in interactive settings is severely undermined by hallucination snowballing: a phenomenon where initial errors amplify across conversational turns, leading to a collapse in coherence. This failure reveals a fundamental vulnerability where models progressively neglect visual grounding in favor of over-relying on polluted textual history. Existing benchmarks are predominantly confined to single-turn VQA, which fail to capture the complex dynamics of error propagation in long-horizon interactions. To address this, we introduce MM-Snowball, the first benchmark for fine-grained diagnosis of hallucination snowballing within dialogues. Extensive evaluation shows that our benchmark poses a significant challenge even to advanced MLLMs and reveals the inefficacy of existing mitigation methods designed for single-turn VQA. To counteract this degradation, we propose Conflict-Aware Visual Rectification (CAVR). This training-free method mitigates snowballing through a synergistic dual-mechanism that refreshes visual grounding at the representation level and rectifies output distributions at the logit level, effectively re-anchoring the model to visual facts. Experiments demonstrate that CAVR achieves state-of-the-art performance, offering a promising path toward more reliable interactive AI. Data and code are available at: https://frenkie-chiang.github.io/MM-Snowball

Mingbo Hong, Yuhang Lu, Nianjin Ye et al.

Estimating homography from an image pair is a fundamental problem in image alignment. Unsupervised learning methods have received increasing attention in this field due to their promising performance and label-free training. However, existing methods do not explicitly consider the problem of plane-induced parallax, which will make the predicted homography compromised on multiple planes. In this work, we propose a novel method HomoGAN to guide unsupervised homography estimation to focus on the dominant plane. First, a multi-scale transformer network is designed to predict homography from the feature pyramids of input images in a coarse-to-fine fashion. Moreover, we propose an unsupervised GAN to impose coplanarity constraint on the predicted homography, which is realized by using a generator to predict a mask of aligned regions, and then a discriminator to check if two masked feature maps are induced by a single homography. To validate the effectiveness of HomoGAN and its components, we conduct extensive experiments on a large-scale dataset, and the results show that our matching error is 22% lower than the previous SOTA method. Code is available at https://github.com/megvii-research/HomoGAN.

Hai Jiang, Haipeng Li, Yuhang Lu et al.

Homography estimation is erroneous in the case of large-baseline due to the low image overlay and limited receptive field. To address it, we propose a progressive estimation strategy by converting large-baseline homography into multiple intermediate ones, cumulatively multiplying these intermediate items can reconstruct the initial homography. Meanwhile, a semi-supervised homography identity loss, which consists of two components: a supervised objective and an unsupervised objective, is introduced. The first supervised loss is acting to optimize intermediate homographies, while the second unsupervised one helps to estimate a large-baseline homography without photometric losses. To validate our method, we propose a large-scale dataset that covers regular and challenging scenes. Experiments show that our method achieves state-of-the-art performance in large-baseline scenes while keeping competitive performance in small-baseline scenes. Code and dataset are available at https://github.com/megvii-research/LBHomo.

Yuhang Lu, Touradj Ebrahimi

Detecting digital face manipulation in images and video has attracted extensive attention due to the potential risk to public trust. To counteract the malicious usage of such techniques, deep learning-based deepfake detection methods have been employed and have exhibited remarkable performance. However, the performance of such detectors is often assessed on related benchmarks that hardly reflect real-world situations. For example, the impact of various image and video processing operations and typical workflow distortions on detection accuracy has not been systematically measured. In this paper, a more reliable assessment framework is proposed to evaluate the performance of learning-based deepfake detectors in more realistic settings. To the best of our acknowledgment, it is the first systematic assessment approach for deepfake detectors that not only reports the general performance under real-world conditions but also quantitatively measures their robustness toward different processing operations. To demonstrate the effectiveness and usage of the framework, extensive experiments and detailed analysis of three popular deepfake detection methods are further presented in this paper. In addition, a stochastic degradation-based data augmentation method driven by realistic processing operations is designed, which significantly improves the robustness of deepfake detectors.

Zewei Xu, Yuhang Lu, Touradj Ebrahimi

Despite the huge success of deep convolutional neural networks in face recognition (FR) tasks, current methods lack explainability for their predictions because of their "black-box" nature. In recent years, studies have been carried out to give an interpretation of the decision of a deep FR system. However, the affinity between the input facial image and the extracted deep features has not been explored. This paper contributes to the problem of explainable face recognition by first conceiving a face reconstruction-based explanation module, which reveals the correspondence between the deep feature and the facial regions. To further interpret the decision of an FR model, a novel visual saliency explanation algorithm has been proposed. It provides insightful explanation by producing visual saliency maps that represent similar and dissimilar regions between input faces. A detailed analysis has been presented for the generated visual explanation to show the effectiveness of the proposed method.

Yuhang Lu, Touradj Ebrahimi

Deep convolutional neural networks have achieved exceptional results on multiple detection and recognition tasks. However, the performance of such detectors are often evaluated in public benchmarks under constrained and non-realistic situations. The impact of conventional distortions and processing operations found in imaging workflows such as compression, noise, and enhancement are not sufficiently studied. Currently, only a few researches have been done to improve the detector robustness to unseen perturbations. This paper proposes a more effective data augmentation scheme based on real-world image degradation process. This novel technique is deployed for deepfake detection tasks and has been evaluated by a more realistic assessment framework. Extensive experiments show that the proposed data augmentation scheme improves generalization ability to unpredictable data distortions and unseen datasets.

Yuhang Lu, Touradj Ebrahimi

The detection of digital face manipulation in video has attracted extensive attention due to the increased risk to public trust. To counteract the malicious usage of such techniques, deep learning-based deepfake detection methods have been developed and have shown impressive results. However, the performance of these detectors is often evaluated using benchmarks that hardly reflect real-world situations. For example, the impact of various video processing operations on detection accuracy has not been systematically assessed. To address this gap, this paper first analyzes numerous real-world influencing factors and typical video processing operations. Then, a more systematic assessment methodology is proposed, which allows for a quantitative evaluation of a detector's robustness under the influence of different processing operations. Moreover, substantial experiments have been carried out on three popular deepfake detectors, which give detailed analyses on the impact of each operation and bring insights to foster future research.

Yuhang Lu, Qi Jiang, Runnan Chen et al.

Zero-shot point cloud segmentation aims to make deep models capable of recognizing novel objects in point cloud that are unseen in the training phase. Recent trends favor the pipeline which transfers knowledge from seen classes with labels to unseen classes without labels. They typically align visual features with semantic features obtained from word embedding by the supervision of seen classes' annotations. However, point cloud contains limited information to fully match with semantic features. In fact, the rich appearance information of images is a natural complement to the textureless point cloud, which is not well explored in previous literature. Motivated by this, we propose a novel multi-modal zero-shot learning method to better utilize the complementary information of point clouds and images for more accurate visual-semantic alignment. Extensive experiments are performed in two popular benchmarks, i.e., SemanticKITTI and nuScenes, and our method outperforms current SOTA methods with 52% and 49% improvement on average for unseen class mIoU, respectively.

Yuhang Lu, Touradj Ebrahimi

Despite the significant progress in face recognition in the past years, they are often treated as "black boxes" and have been criticized for lacking explainability. It becomes increasingly important to understand the characteristics and decisions of deep face recognition systems to make them more acceptable to the public. Explainable face recognition (XFR) refers to the problem of interpreting why the recognition model matches a probe face with one identity over others. Recent studies have explored use of visual saliency maps as an explanation, but they often lack a deeper analysis in the context of face recognition. This paper starts by proposing a rigorous definition of explainable face recognition (XFR) which focuses on the decision-making process of the deep recognition model. Following the new definition, a similarity-based RISE algorithm (S-RISE) is then introduced to produce high-quality visual saliency maps. Furthermore, an evaluation approach is proposed to systematically validate the reliability and accuracy of general visual saliency-based XFR methods.

Yuhang Lu, Touradj Ebrahimi

Cross-resolution face recognition has become a challenging problem for modern deep face recognition systems. It aims at matching a low-resolution probe image with high-resolution gallery images registered in a database. Existing methods mainly leverage prior information from high-resolution images by either reconstructing facial details with super-resolution techniques or learning a unified feature space. To address this challenge, this paper proposes a new approach that enforces the network to focus on the discriminative information stored in the low-frequency components of a low-resolution image. A cross-resolution knowledge distillation paradigm is first employed as the learning framework. Then, an identity-preserving network, WaveResNet, and a wavelet similarity loss are designed to capture low-frequency details and boost performance. Finally, an image degradation model is conceived to simulate more realistic low-resolution training data. Consequently, extensive experimental results show that the proposed method consistently outperforms the baseline model and other state-of-the-art methods across a variety of image resolutions.

Yuhang Lu, Ruizhi Luo, Touradj Ebrahimi

Deep convolutional neural networks have shown remarkable results on multiple detection tasks. Despite the significant progress, the performance of such detectors are often assessed in public benchmarks under non-realistic conditions. Specifically, impact of conventional distortions and processing operations such as compression, noise, and enhancement are not sufficiently studied. This paper proposes a rigorous framework to assess performance of learning-based detectors in more realistic situations. An illustrative example is shown under deepfake detection context. Inspired by the assessment results, a data augmentation strategy based on natural image degradation process is designed, which significantly improves the generalization ability of two deepfake detectors.

Yuhang Lu, Xinyi Wu, Zhenyao Wu et al.

Few-shot segmentation (FSS) expects models trained on base classes to work on novel classes with the help of a few support images. However, when there exists a domain gap between the base and novel classes, the state-of-the-art FSS methods may even fail to segment simple objects. To improve their performance on unseen domains, we propose to transductively fine-tune the base model on a set of query images under the few-shot setting, where the core idea is to implicitly guide the segmentation of query images using support labels. Although different images are not directly comparable, their class-wise prototypes are desired to be aligned in the feature space. By aligning query and support prototypes with an uncertainty-aware contrastive loss, and using a supervised cross-entropy loss and an unsupervised boundary loss as regularizations, our method could generalize the base model to the target domain without additional labels. We conduct extensive experiments under various cross-domain settings of natural, remote sensing, and medical images. The results show that our method could consistently and significantly improve the performance of prototypical FSS models in all cross-domain tasks.

Ziyang Yan, Wenzhen Dong, Yihua Shao et al.

End-to-end autonomous driving with vision-only is not only more cost-effective compared to LiDAR-vision fusion but also more reliable than traditional methods. To achieve a economical and robust purely visual autonomous driving system, we propose RenderWorld, a vision-only end-to-end autonomous driving framework, which generates 3D occupancy labels using a self-supervised gaussian-based Img2Occ Module, then encodes the labels by AM-VAE, and uses world model for forecasting and planning. RenderWorld employs Gaussian Splatting to represent 3D scenes and render 2D images greatly improves segmentation accuracy and reduces GPU memory consumption compared with NeRF-based methods. By applying AM-VAE to encode air and non-air separately, RenderWorld achieves more fine-grained scene element representation, leading to state-of-the-art performance in both 4D occupancy forecasting and motion planning from autoregressive world model.

Kuo Liang, Yuhang Lu, Jianming Mao et al.

Large-scale optimization is a key backbone of modern business decision-making. However, building these models is often labor-intensive and time-consuming. We address this by proposing LEAN-LLM-OPT, a LightwEight AgeNtic workflow construction framework for LLM-assisted large-scale OPTimization auto-formulation. LEAN-LLM-OPT takes as input a problem description together with associated datasets and orchestrates a team of LLM agents to produce an optimization formulation. Specifically, upon receiving a query, two upstream LLM agents dynamically construct a workflow that specifies, step-by-step, how optimization models for similar problems can be formulated. A downstream LLM agent then follows this workflow to generate the final output. Leveraging LLMs' text-processing capabilities and common modeling practices, the workflow decomposes the modeling task into a sequence of structured sub-tasks and offloads mechanical data-handling operations to auxiliary tools. This design alleviates the downstream agent's burden related to planning and data handling, allowing it to focus on the most challenging components that cannot be readily standardized. Extensive simulations show that LEAN-LLM-OPT, instantiated with GPT-4.1 and the open source gpt-oss-20B, achieves strong performance on large-scale optimization modeling tasks and is competitive with state-of-the-art approaches. In addition, in a Singapore Airlines choice-based revenue management use case, LEAN-LLM-OPT demonstrates practical value by achieving leading performance across a range of scenarios. Along the way, we introduce Large-Scale-OR and Air-NRM, the first comprehensive benchmarks for large-scale optimization auto-formulation. The code and data of this work is available at https://github.com/CoraLiang01/lean-llm-opt.

Yuhang Lu, Yichen Yao, Jiadong Tu et al.

Large Vision-Language Models (LVLMs) have recently garnered significant attention, with many efforts aimed at harnessing their general knowledge to enhance the interpretability and robustness of autonomous driving models. However, LVLMs typically rely on large, general-purpose datasets and lack the specialized expertise required for professional and safe driving. Existing vision-language driving datasets focus primarily on scene understanding and decision-making, without providing explicit guidance on traffic rules and driving skills, which are critical aspects directly related to driving safety. To bridge this gap, we propose IDKB, a large-scale dataset containing over one million data items collected from various countries, including driving handbooks, theory test data, and simulated road test data. Much like the process of obtaining a driver's license, IDKB encompasses nearly all the explicit knowledge needed for driving from theory to practice. In particular, we conducted comprehensive tests on 15 LVLMs using IDKB to assess their reliability in the context of autonomous driving and provided extensive analysis. We also fine-tuned popular models, achieving notable performance improvements, which further validate the significance of our dataset. The project page can be found at: \url{https://4dvlab.github.io/project_page/idkb.html}

Yuhang Lu, Zewei Xu, Touradj Ebrahimi

Over recent years, deep convolutional neural networks have significantly advanced the field of face recognition techniques for both verification and identification purposes. Despite the impressive accuracy, these neural networks are often criticized for lacking explainability. There is a growing demand for understanding the decision-making process of AI-based face recognition systems. Some studies have investigated the use of visual saliency maps as explanations, but they have predominantly focused on the specific face verification case. The discussion on more general face recognition scenarios and the corresponding evaluation methodology for these explanations have long been absent in current research. Therefore, this manuscript conceives a comprehensive explanation framework for face recognition tasks. Firstly, an exhaustive definition of visual saliency map-based explanations for AI-based face recognition systems is provided, taking into account the two most common recognition situations individually, i.e., face verification and identification. Secondly, a new model-agnostic explanation method named CorrRISE is proposed to produce saliency maps, which reveal both the similar and dissimilar regions between any given face images. Subsequently, the explanation framework conceives a new evaluation methodology that offers quantitative measurement and comparison of the performance of general visual saliency explanation methods in face recognition. Consequently, extensive experiments are carried out on multiple verification and identification scenarios. The results showcase that CorrRISE generates insightful saliency maps and demonstrates superior performance, particularly in similarity maps in comparison with the state-of-the-art explanation approaches.

Enhui Ma, Jiahuan Zhang, Guantian Zheng et al.

Multimodal Large Language Models (MLLMs) are rapidly becoming the intelligence brain of end-to-end autonomous driving systems. A key challenge is to assess whether MLLMs can truly understand and follow complex real-world traffic rules. However, existing benchmarks mainly focus on single-rule scenarios like traffic sign recognition, neglecting the complexity of multi-rule concurrency and conflicts in real driving. Consequently, models perform well on simple tasks but often fail or violate rules in real world complex situations. To bridge this gap, we propose DriveCombo, a text and vision-based benchmark for compositional traffic rule reasoning. Inspired by human drivers' cognitive development, we propose a systematic Five-Level Cognitive Ladder that evaluates reasoning from single-rule understanding to multi-rule integration and conflict resolution, enabling quantitative assessment across cognitive stages. We further propose a Rule2Scene Agent that maps language-based traffic rules to dynamic driving scenes through rule crafting and scene generation, enabling scene-level traffic rule visual reasoning. Evaluations of 14 mainstream MLLMs reveal performance drops as task complexity grows, particularly during rule conflicts. After splitting the dataset and fine-tuning on the training set, we further observe substantial improvements in both traffic rule reasoning and downstream planning capabilities. These results highlight the effectiveness of DriveCombo in advancing compliant and intelligent autonomous driving systems.

Zhanpeng Wang, Shuting Cao, Yuhang Lu et al.

The Dual Diffusion Implicit Bridge (DDIB) is an emerging image-to-image (I2I) translation method that preserves cycle consistency while achieving strong flexibility. It links two independently trained diffusion models (DMs) in the source and target domains by first adding noise to a source image to obtain a latent code, then denoising it in the target domain to generate the translated image. However, this method faces two key challenges: (1) low translation efficiency, and (2) translation trajectory deviations caused by mismatched latent distributions. To address these issues, we propose a novel I2I translation framework, OT-ALD, grounded in optimal transport (OT) theory, which retains the strengths of DDIB-based approach. Specifically, we compute an OT map from the latent distribution of the source domain to that of the target domain, and use the mapped distribution as the starting point for the reverse diffusion process in the target domain. Our error analysis confirms that OT-ALD eliminates latent distribution mismatches. Moreover, OT-ALD effectively balances faster image translation with improved image quality. Experiments on four translation tasks across three high-resolution datasets show that OT-ALD improves sampling efficiency by 20.29% and reduces the FID score by 2.6 on average compared to the top-performing baseline models.

Yuhang Lu, Xinge Zhu, Tai Wang et al.

Occupancy prediction has increasingly garnered attention in recent years for its fine-grained understanding of 3D scenes. Traditional approaches typically rely on dense, regular grid representations, which often leads to excessive computational demands and a loss of spatial details for small objects. This paper introduces OctreeOcc, an innovative 3D occupancy prediction framework that leverages the octree representation to adaptively capture valuable information in 3D, offering variable granularity to accommodate object shapes and semantic regions of varying sizes and complexities. In particular, we incorporate image semantic information to improve the accuracy of initial octree structures and design an effective rectification mechanism to refine the octree structure iteratively. Our extensive evaluations show that OctreeOcc not only surpasses state-of-the-art methods in occupancy prediction, but also achieves a 15%-24% reduction in computational overhead compared to dense-grid-based methods.

Yuhang Lu, Touradj Ebrahimi

Over the past years, image generation and manipulation have achieved remarkable progress due to the rapid development of generative AI based on deep learning. Recent studies have devoted significant efforts to address the problem of face image manipulation caused by deepfake techniques. However, the problem of detecting purely synthesized face images has been explored to a lesser extent. In particular, the recent popular Diffusion Models (DMs) have shown remarkable success in image synthesis. Existing detectors struggle to generalize between synthesized images created by different generative models. In this work, a comprehensive benchmark including human face images produced by Generative Adversarial Networks (GANs) and a variety of DMs has been established to evaluate both the generalization ability and robustness of state-of-the-art detectors. Then, the forgery traces introduced by different generative models have been analyzed in the frequency domain to draw various insights. The paper further demonstrates that a detector trained with frequency representation can generalize well to other unseen generative models.

Yuhang Lu, Zewei Xu, Touradj Ebrahimi

Recent years have witnessed significant advancement in face recognition (FR) techniques, with their applications widely spread in people's lives and security-sensitive areas. There is a growing need for reliable interpretations of decisions of such systems. Existing studies relying on various mechanisms have investigated the usage of saliency maps as an explanation approach, but suffer from different limitations. This paper first explores the spatial relationship between face image and its deep representation via gradient backpropagation. Then a new explanation approach FGGB has been conceived, which provides precise and insightful similarity and dissimilarity saliency maps to explain the "Accept" and "Reject" decision of an FR system. Extensive visual presentation and quantitative measurement have shown that FGGB achieves superior performance in both similarity and dissimilarity maps when compared to current state-of-the-art explainable face verification approaches.

Yuhang Lu, Jiadong Tu, Yuexin Ma et al.

End-to-end autonomous driving has emerged as a promising approach to unify perception, prediction, and planning within a single framework, reducing information loss and improving adaptability. However, existing methods often rely on fixed and sparse trajectory supervision, limiting their ability to capture the hierarchical reasoning process that human drivers naturally employ. To bridge this gap, we propose ReAL-AD, a Reasoning-Augmented Learning framework that structures decision-making in autonomous driving based on the three-tier human cognitive model: Driving Strategy, Driving Decision, and Driving Operation, where Vision-Language Models (VLMs) are incorporated to enhance situational awareness and structured reasoning across these levels. Specifically, we introduce: (1) the Strategic Reasoning Injector, which formulates high-level driving strategies by interpreting complex traffic contexts from VLM-generated insights; (2) the Tactical Reasoning Integrator, which refines strategic intent into interpretable tactical choices such as lane changes, overtaking, and speed adjustments; and (3) the Hierarchical Trajectory Decoder, which progressively translates tactical decisions into precise control actions for smooth and human-like trajectory execution. Extensive evaluations show that integrating our framework improves planning accuracy and safety by over 30%, making end-to-end autonomous driving more interpretable and aligned with human-like hierarchical reasoning. The project page can be found at: \href{https://4dvlab.github.io/project_page/realad}{\texttt{4dvlab.github.io/project\_page/realad}}

Nikolay Stanishev, Yuhang Lu, Touradj Ebrahimi

Pose-invariant face recognition has become a challenging problem for modern AI-based face recognition systems. It aims at matching a profile face captured in the wild with a frontal face registered in a database. Existing methods perform face frontalization via either generative models or learning a pose robust feature representation. In this paper, a new method is presented to perform face frontalization and recognition within the feature space. First, a novel feature space pose frontalization module (FSPFM) is proposed to transform profile images with arbitrary angles into frontal counterparts. Second, a new training paradigm is proposed to maximize the potential of FSPFM and boost its performance. The latter consists of a pre-training and an attention-guided fine-tuning stage. Moreover, extensive experiments have been conducted on five popular face recognition benchmarks. Results show that not only our method outperforms the state-of-the-art in the pose-invariant face recognition task but also maintains superior performance in other standard scenarios.

Yudan Song, Yuecen Wei, Yuhang Lu et al.

Graph representation learning has become a mainstream method for fraud detection due to its strong expressive power, which focuses on enhancing node representations through improved neighborhood knowledge capture. However, the focus on local interactions leads to imbalanced transmission of global topological information and increased risk of node-specific information being overwhelmed during aggregation due to the imbalance between fraud and benign nodes. In this paper, we first summarize the impact of topology and class imbalance on downstream tasks in GNN-based fraud detection, as the problem of imbalanced supervisory messages is caused by fraudsters' topological behavior obfuscation and identity feature concealment. Based on statistical validation, we propose a novel dual-view graph representation learning method to mitigate Message imbalance in Fraud Detection (MimbFD). Specifically, we design a topological message reachability module for high-quality node representation learning to penetrate fraudsters' camouflage and alleviate insufficient propagation. Then, we introduce a local confounding debiasing module to adjust node representations, enhancing the stable association between node representations and labels to balance the influence of different classes. Finally, we conducted experiments on three public fraud datasets, and the results demonstrate that MimbFD exhibits outstanding performance in fraud detection.

Yuhang Lu, Zewei Xu, Touradj Ebrahimi

In the past years, deep convolutional neural networks have been pushing the frontier of face recognition (FR) techniques in both verification and identification scenarios. Despite the high accuracy, they are often criticized for lacking explainability. There has been an increasing demand for understanding the decision-making process of deep face recognition systems. Recent studies have investigated the usage of visual saliency maps as an explanation, but they often lack a discussion and analysis in the context of face recognition. This paper concentrates on explainable face verification tasks and conceives a new explanation framework. Firstly, a definition of the saliency-based explanation method is provided, which focuses on the decisions made by the deep FR model. Secondly, a new model-agnostic explanation method named CorrRISE is proposed to produce saliency maps, which reveal both the similar and dissimilar regions of any given pair of face images. Then, an evaluation methodology is designed to measure the performance of general visual saliency explanation methods in face verification. Finally, substantial visual and quantitative results have shown that the proposed CorrRISE method demonstrates promising results in comparison with other state-of-the-art explainable face verification approaches.

Xinyi Wu, Zhenyao Wu, Yuhang Lu et al.

In this paper, we tackle the problem of one-shot unsupervised domain adaptation (OSUDA) for semantic segmentation where the segmentors only see one unlabeled target image during training. In this case, traditional unsupervised domain adaptation models usually fail since they cannot adapt to the target domain with over-fitting to one (or few) target samples. To address this problem, existing OSUDA methods usually integrate a style-transfer module to perform domain randomization based on the unlabeled target sample, with which multiple domains around the target sample can be explored during training. However, such a style-transfer module relies on an additional set of images as style reference for pre-training and also increases the memory demand for domain adaptation. Here we propose a new OSUDA method that can effectively relieve such computational burden. Specifically, we integrate several style-mixing layers into the segmentor which play the role of style-transfer module to stylize the source images without introducing any learned parameters. Moreover, we propose a patchwise prototypical matching (PPM) method to weighted consider the importance of source pixels during the supervised training to relieve the negative adaptation. Experimental results show that our method achieves new state-of-the-art performance on two commonly used benchmarks for domain adaptive semantic segmentation under the one-shot setting and is more efficient than all comparison approaches.

Yuhang Lu, Evgeniy Upenik, Touradj Ebrahimi

Deepfakes are becoming increasingly popular in both good faith applications such as in entertainment and maliciously intended manipulations such as in image and video forgery. Primarily motivated by the latter, a large number of deepfake detectors have been proposed recently in order to identify such content. While the performance of such detectors still need further improvements, they are often assessed in simple if not trivial scenarios. In particular, the impact of benign processing operations such as transcoding, denoising, resizing and enhancement are not sufficiently studied. This paper proposes a more rigorous and systematic framework to assess the performance of deepfake detectors in more realistic situations. It quantitatively measures how and to which extent each benign processing approach impacts a state-of-the-art deepfake detection method. By illustrating it in a popular deepfake detector, our benchmark proposes a framework to assess robustness of detectors and provides valuable insights to design more efficient deepfake detectors.

Yuhang Lu, Kang Zheng, Weijian Li et al.

Accurate segmentation of anatomical structures is vital for medical image analysis. The state-of-the-art accuracy is typically achieved by supervised learning methods, where gathering the requisite expert-labeled image annotations in a scalable manner remains a main obstacle. Therefore, annotation-efficient methods that permit to produce accurate anatomical structure segmentation are highly desirable. In this work, we present Contour Transformer Network (CTN), a one-shot anatomy segmentation method with a naturally built-in human-in-the-loop mechanism. We formulate anatomy segmentation as a contour evolution process and model the evolution behavior by graph convolutional networks (GCNs). Training the CTN model requires only one labeled image exemplar and leverages additional unlabeled data through newly introduced loss functions that measure the global shape and appearance consistency of contours. On segmentation tasks of four different anatomies, we demonstrate that our one-shot learning method significantly outperforms non-learning-based methods and performs competitively to the state-of-the-art fully supervised deep learning methods. With minimal human-in-the-loop editing feedback, the segmentation performance can be further improved to surpass the fully supervised methods.

Yuhang Lu, Weijian Li, Kang Zheng et al.

Accurate segmentation of critical anatomical structures is at the core of medical image analysis. The main bottleneck lies in gathering the requisite expert-labeled image annotations in a scalable manner. Methods that permit to produce accurate anatomical structure segmentation without using a large amount of fully annotated training images are highly desirable. In this work, we propose a novel contribution of Contour Transformer Network (CTN), a one-shot anatomy segmentor including a naturally built-in human-in-the-loop mechanism. Segmentation is formulated by learning a contour evolution behavior process based on graph convolutional networks (GCNs). Training of our CTN model requires only one labeled image exemplar and leverages additional unlabeled data through newly introduced loss functions that measure the global shape and appearance consistency of contours. We demonstrate that our one-shot learning method significantly outperforms non-learning-based methods and performs competitively to the state-of-the-art fully supervised deep learning approaches. With minimal human-in-the-loop editing feedback, the segmentation performance can be further improved and tailored towards the observer desired outcomes. This can facilitate the clinician designed imaging-based biomarker assessments (to support personalized quantitative clinical diagnosis) and outperforms fully supervised baselines.

Weijian Li, Yuhang Lu, Kang Zheng et al.

Image landmark detection aims to automatically identify the locations of predefined fiducial points. Despite recent success in this field, higher-ordered structural modeling to capture implicit or explicit relationships among anatomical landmarks has not been adequately exploited. In this work, we present a new topology-adapting deep graph learning approach for accurate anatomical facial and medical (e.g., hand, pelvis) landmark detection. The proposed method constructs graph signals leveraging both local image features and global shape features. The adaptive graph topology naturally explores and lands on task-specific structures which are learned end-to-end with two Graph Convolutional Networks (GCNs). Extensive experiments are conducted on three public facial image datasets (WFLW, 300W, and COFW-68) as well as three real-world X-ray medical datasets (Cephalometric (public), Hand and Pelvis). Quantitative results comparing with the previous state-of-the-art approaches across all studied datasets indicating the superior performance in both robustness and accuracy. Qualitative visualizations of the learned graph topologies demonstrate a physically plausible connectivity laying behind the landmarks.

Yanting Pei, Yaping Huang, Qi Zou et al.

Hazy images are common in real scenarios and many dehazing methods have been developed to automatically remove the haze from images. Typically, the goal of image dehazing is to produce clearer images from which human vision can better identify the object and structural details present in the images. When the ground-truth haze-free image is available for a hazy image, quantitative evaluation of image dehazing is usually based on objective metrics, such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity (SSIM). However, in many applications, large-scale images are collected not for visual examination by human. Instead, they are used for many high-level vision tasks, such as automatic classification, recognition and categorization. One fundamental problem here is whether various dehazing methods can produce clearer images that can help improve the performance of the high-level tasks. In this paper, we empirically study this problem in the important task of image classification by using both synthetic and real hazy image datasets. From the experimental results, we find that the existing image-dehazing methods cannot improve much the image-classification performance and sometimes even reduce the image-classification performance.

Jun Zhou, Yuhang Lu, Kang Zheng et al.

The surfaces of many cultural heritage objects were embellished with various patterns, especially curve patterns. In practice, most of the unearthed cultural heritage objects are highly fragmented, e.g., sherds of potteries or vessels, and each of them only shows a very small portion of the underlying full design, with noise and deformations. The goal of this paper is to address the challenging problem of automatically identifying the underlying full design of curve patterns from such a sherd. Specifically, we formulate this problem as template matching: curve structure segmented from the sherd is matched to each location with each possible orientation of each known full design. In this paper, we propose a new two-stage matching algorithm, with a different matching cost in each stage. In Stage 1, we use a traditional template matching, which is highly computationally efficient, over the whole search space and identify a small set of candidate matchings. In Stage 2, we derive a new matching cost by training a dual-source Convolutional Neural Network (CNN) and apply it to re-rank the candidate matchings identified in Stage 1. We collect 600 pottery sherds with 98 full designs from the Woodland Period in Southeastern North America for experiments and the performance of the proposed algorithm is very competitive.

Jing Yu, Yuhang Lu, Zengchang Qin et al.

Cross-modal information retrieval aims to find heterogeneous data of various modalities from a given query of one modality. The main challenge is to map different modalities into a common semantic space, in which distance between concepts in different modalities can be well modeled. For cross-modal information retrieval between images and texts, existing work mostly uses off-the-shelf Convolutional Neural Network (CNN) for image feature extraction. For texts, word-level features such as bag-of-words or word2vec are employed to build deep learning models to represent texts. Besides word-level semantics, the semantic relations between words are also informative but less explored. In this paper, we model texts by graphs using similarity measure based on word2vec. A dual-path neural network model is proposed for couple feature learning in cross-modal information retrieval. One path utilizes Graph Convolutional Network (GCN) for text modeling based on graph representations. The other path uses a neural network with layers of nonlinearities for image modeling based on off-the-shelf features. The model is trained by a pairwise similarity loss function to maximize the similarity of relevant text-image pairs and minimize the similarity of irrelevant pairs. Experimental results show that the proposed model outperforms the state-of-the-art methods significantly, with 17% improvement on accuracy for the best case.

Yuhang Lu, Jun Zhou, Jing Wang et al.

Motivated by the important archaeological application of exploring cultural heritage objects, in this paper we study the challenging problem of automatically segmenting curve structures that are very weakly stamped or carved on an object surface in the form of a highly noisy depth map. Different from most classical low-level image segmentation methods that are known to be very sensitive to the noise and occlusions, we propose a new supervised learning algorithm based on Convolutional Neural Network (CNN) to implicitly learn and utilize more curve geometry and pattern information for addressing this challenging problem. More specifically, we first propose a Fully Convolutional Network (FCN) to estimate the skeleton of curve structures and at each skeleton pixel, a scale value is estimated to reflect the local curve width. Then we propose a dense prediction network to refine the estimated curve skeletons. Based on the estimated scale values, we finally develop an adaptive thresholding algorithm to achieve the final segmentation of curve structures. In the experiment, we validate the performance of the proposed method on a dataset of depth images scanned from unearthed pottery sherds dating to the Woodland period of Southeastern North America.

Yuhang Lu, Youchuan Wan, Gang Li

Unsupervised evaluation of segmentation quality is a crucial step in image segmentation applications. Previous unsupervised evaluation methods usually lacked the adaptability to multi-scale segmentation. A scale-constrained evaluation method that evaluates segmentation quality according to the specified target scale is proposed in this paper. First, regional saliency and merging cost are employed to describe intra-region homogeneity and inter-region heterogeneity, respectively. Subsequently, both of them are standardized into equivalent spectral distances of a predefined region. Finally, by analyzing the relationship between image characteristics and segmentation quality, we establish the evaluation model. Experimental results show that the proposed method outperforms four commonly used unsupervised methods in multi-scale evaluation tasks.