Haohang Xu

Shuangrui Ding, Rui Qian, Haohang Xu et al.

In this paper, we propose a simple yet effective approach for self-supervised video object segmentation (VOS). Our key insight is that the inherent structural dependencies present in DINO-pretrained Transformers can be leveraged to establish robust spatio-temporal correspondences in videos. Furthermore, simple clustering on this correspondence cue is sufficient to yield competitive segmentation results. Previous self-supervised VOS techniques majorly resort to auxiliary modalities or utilize iterative slot attention to assist in object discovery, which restricts their general applicability and imposes higher computational requirements. To deal with these challenges, we develop a simplified architecture that capitalizes on the emerging objectness from DINO-pretrained Transformers, bypassing the need for additional modalities or slot attention. Specifically, we first introduce a single spatio-temporal Transformer block to process the frame-wise DINO features and establish spatio-temporal dependencies in the form of self-attention. Subsequently, utilizing these attention maps, we implement hierarchical clustering to generate object segmentation masks. To train the spatio-temporal block in a fully self-supervised manner, we employ semantic and dynamic motion consistency coupled with entropy normalization. Our method demonstrates state-of-the-art performance across multiple unsupervised VOS benchmarks and particularly excels in complex real-world multi-object video segmentation tasks such as DAVIS-17-Unsupervised and YouTube-VIS-19. The code and model checkpoints will be released at https://github.com/shvdiwnkozbw/SSL-UVOS.

Haohang Xu, Shuangrui Ding, Manqi Zhao et al.

Deep neural networks are capable of learning powerful representations to tackle complex vision tasks but expose undesirable properties like the over-fitting issue. To this end, regularization techniques like image augmentation are necessary for deep neural networks to generalize well. Nevertheless, most prevalent image augmentation recipes confine themselves to off-the-shelf linear transformations like scale, flip, and colorjitter. Due to their hand-crafted property, these augmentations are insufficient to generate truly hard augmented examples. In this paper, we propose a novel perspective of augmentation to regularize the training process. Inspired by the recent success of applying masked image modeling to self-supervised learning, we adopt the self-supervised masked autoencoder to generate the distorted view of the input images. We show that utilizing such model-based nonlinear transformation as data augmentation can improve high-level recognition tasks. We term the proposed method as \textbf{M}ask-\textbf{R}econstruct \textbf{A}ugmentation (MRA). The extensive experiments on various image classification benchmarks verify the effectiveness of the proposed augmentation. Specifically, MRA consistently enhances the performance on supervised, semi-supervised as well as few-shot classification.

Haohang Xu, Lin Liu, Zhibo Zhang et al.

Diffusion-based image editing models have achieved significant progress in real world applications. However, conventional models typically rely on natural language prompts, which often lack the precision required to localize target objects. Consequently, these models struggle to maintain background consistency due to their global image regeneration paradigm. Recognizing that visual cues provide an intuitive means for users to highlight specific areas of interest, we utilize bounding boxes as guidance to explicitly define the editing target. This approach ensures that the diffusion model can accurately localize the target while preserving background consistency. To achieve this, we propose FineEdit, a multi-level bounding box injection method that enables the model to utilize spatial conditions more effectively. To support this high precision guidance, we present FineEdit-1.2M, a large scale, fine-grained dataset comprising 1.2 million image editing pairs with precise bounding box annotations. Furthermore, we construct a comprehensive benchmark, termed FineEdit-Bench, which includes 1,000 images across 10 subjects to effectively evaluate region based editing capabilities. Evaluations on FineEdit-Bench demonstrate that our model significantly outperforms state-of-the-art open-source models (e.g., Qwen-Image-Edit and LongCat-Image-Edit) in instruction compliance and background preservation. Further assessments on open benchmarks (GEdit and ImgEdit Bench) confirm its superior generalization and robustness.

Lin Liu, Zhihan Xiao, Haohang Xu et al.

Video editing has recently achieved remarkable progress with diffusion-based generative models, enabling diverse object-level manipulations from natural language instructions. However, existing methods often struggle under occlusion, viewpoint changes, and fast object motion, where unreliable visual observations lead to inaccurate localization, temporal flickering, and inconsistent edits. In this work, we identify the absence of reliable visual anchors as a fundamental bottleneck in occlusion-robust video editing. To address this issue, we propose an occlusion-aware physics-semantic keyframe selection framework that automatically identifies an optimal anchor frame for downstream editing. Specifically, our method evaluates candidate frames from three complementary perspectives: structural completeness for avoiding truncated observations, cycle-consistent tracking stability for measuring physical reliability, and vision-language-based attribute visibility for ensuring semantic clarity. The selected keyframe is then propagated through bidirectional tracking to generate dense spatiotemporal masks, which are used as auxiliary supervision for a diffusion-based video editing backbone. By transforming occlusion handling from explicit reconstruction into reliable anchor selection, our framework enables precise and temporally consistent editing without requiring manual annotations. Extensive experiments on challenging video editing benchmarks demonstrate the effectiveness and high-quality performance of our method.

Peisen Zhao, Xiaopeng Zhang, Mingxing Xu et al.

While Multimodal Large Language Models (MLLMs) have experienced rapid advancements, their visual encoders frequently remain a performance bottleneck. Conventional CLIP-based encoders struggle with dense spatial tasks due to the loss of visual details caused by low-resolution pretraining and the reliance on noisy, coarse web-crawled image-text pairs. To overcome these limitations, we introduce FineViT, a novel vision encoder specifically designed to unlock fine-grained perception. By replacing coarse web data with dense recaptions, we systematically mitigate information loss through a progressive training paradigm.: first, the encoder is trained from scratch at a high native resolution on billions of global recaptioned image-text pairs, establishing a robust, detail rich semantic foundation. Subsequently, we further enhance its local perception through LLM alignment, utilizing our curated FineCap-450M dataset that comprises over $450$ million high quality local captions. Extensive experiments validate the effectiveness of the progressive strategy. FineViT achieves state-of-the-art zero-shot recognition and retrieval performance, especially in long-context retrieval, and consistently outperforms multimodal visual encoders such as SigLIP2 and Qwen-ViT when integrated into MLLMs. We hope FineViT could serve as a powerful new baseline for fine-grained visual perception.

Haohang Xu, Chengjie Liu, Qihang Wang et al.

Large Language Model (LLM) exhibits great potential in designing of analog integrated circuits (IC) because of its excellence in abstraction and generalization for knowledge. However, further development of LLM-based analog ICs heavily relies on textual description of analog ICs, while existing analog ICs are mostly illustrated in image-based circuit diagrams rather than text-based netlists. Converting circuit diagrams to netlists help LLMs to enrich the knowledge of analog IC. Nevertheless, previously proposed conversion frameworks face challenges in further application because of limited support of image styles and circuit elements. Up to now, it still remains a challenging task to effectively convert complex circuit diagrams into netlists. To this end, this paper constructs and opensources a new dataset with rich styles of circuit diagrams as well as balanced distribution of simple and complex analog ICs. And a hybrid framework, named Image2Net, is proposed for practical conversion from circuit diagrams to netlists. The netlist edit distance (NED) is also introduced to precisely assess the difference between the converted netlists and ground truth. Based on our benchmark, Image2Net achieves 80.77\% successful rate, which is 34.62\%-45.19\% higher than previous works. Specifically, the proposed work shows 0.116 averaged NED, which is 62.1\%-69.6\% lower than state-of-the-arts.

Shuangrui Ding, Maomao Li, Tianyu Yang et al.

In light of the success of contrastive learning in the image domain, current self-supervised video representation learning methods usually employ contrastive loss to facilitate video representation learning. When naively pulling two augmented views of a video closer, the model however tends to learn the common static background as a shortcut but fails to capture the motion information, a phenomenon dubbed as background bias. Such bias makes the model suffer from weak generalization ability, leading to worse performance on downstream tasks such as action recognition. To alleviate such bias, we propose \textbf{F}oreground-b\textbf{a}ckground \textbf{Me}rging (FAME) to deliberately compose the moving foreground region of the selected video onto the static background of others. Specifically, without any off-the-shelf detector, we extract the moving foreground out of background regions via the frame difference and color statistics, and shuffle the background regions among the videos. By leveraging the semantic consistency between the original clips and the fused ones, the model focuses more on the motion patterns and is debiased from the background shortcut. Extensive experiments demonstrate that FAME can effectively resist background cheating and thus achieve the state-of-the-art performance on downstream tasks across UCF101, HMDB51, and Diving48 datasets. The code and configurations are released at https://github.com/Mark12Ding/FAME.

Haohang Xu, Jiemin Fang, Xiaopeng Zhang et al.

Recent advances in self-supervised learning have experienced remarkable progress, especially for contrastive learning based methods, which regard each image as well as its augmentations as an individual class and try to distinguish them from all other images. However, due to the large quantity of exemplars, this kind of pretext task intrinsically suffers from slow convergence and is hard for optimization. This is especially true for small-scale models, in which we find the performance drops dramatically comparing with its supervised counterpart. In this paper, we propose a simple but effective distillation strategy for unsupervised learning. The highlight is that the relationship among similar samples counts and can be seamlessly transferred to the student to boost the performance. Our method, termed as BINGO, which is short for Bag of InstaNces aGgregatiOn, targets at transferring the relationship learned by the teacher to the student. Here bag of instances indicates a set of similar samples constructed by the teacher and are grouped within a bag, and the goal of distillation is to aggregate compact representations over the student with respect to instances in a bag. Notably, BINGO achieves new state-of-the-art performance on small-scale models, i.e., 65.5% and 68.9% top-1 accuracies with linear evaluation on ImageNet, using ResNet-18 and ResNet-34 as the backbones respectively, surpassing baselines (52.5% and 57.4% top-1 accuracies) by a significant margin. The code is available at https://github.com/haohang96/bingo.

Chang Zou, Evelyn Zhang, Runlin Guo et al.

Diffusion Transformers (DiT) have become the dominant methods in image and video generation yet still suffer substantial computational costs. As an effective approach for DiT acceleration, feature caching methods are designed to cache the features of DiT in previous timesteps and reuse them in the next timesteps, allowing us to skip the computation in the next timesteps. Among them, token-wise feature caching has been introduced to perform different caching ratios for different tokens in DiTs, aiming to skip the computation for unimportant tokens while still computing the important ones. In this paper, we propose to carefully check the effectiveness in token-wise feature caching with the following two questions: (1) Is it really necessary to compute the so-called "important" tokens in each step? (2) Are so-called important tokens really important? Surprisingly, this paper gives some counter-intuition answers, demonstrating that consistently computing the selected ``important tokens'' in all steps is not necessary. The selection of the so-called ``important tokens'' is often ineffective, and even sometimes shows inferior performance than random selection. Based on these observations, this paper introduces dual feature caching referred to as DuCa, which performs aggressive caching strategy and conservative caching strategy iteratively and selects the tokens for computing randomly. Extensive experimental results demonstrate the effectiveness of our method in DiT, PixArt, FLUX, and OpenSora, demonstrating significant improvements than the previous token-wise feature caching.

Lei Ke, Haohang Xu, Xuefei Ning et al.

Diffusion models have achieved significant progress in both image and video generation while still suffering from huge computation costs. As an effective solution, flow matching aims to reflow the diffusion process of diffusion models into a straight line for a few-step and even one-step generation. However, in this paper, we suggest that the original training pipeline of flow matching is not optimal and introduce two techniques to improve it. Firstly, we introduce progressive reflow, which progressively reflows the diffusion models in local timesteps until the whole diffusion progresses, reducing the difficulty of flow matching. Second, we introduce aligned v-prediction, which highlights the importance of direction matching in flow matching over magnitude matching. Experimental results on SDv1.5 and SDXL demonstrate the effectiveness of our method, for example, conducting on SDv1.5 achieves an FID of 10.70 on MSCOCO2014 validation set with only 4 sampling steps, close to our teacher model (32 DDIM steps, FID = 10.05).

Haohang Xu, Longyu Chen, Yichen Zhang et al.

While diffusion-based generative models have made significant strides in visual content creation, conventional approaches face computational challenges, especially for high-resolution images, as they denoise the entire image from noisy inputs. This contrasts with signal processing techniques, such as Fourier and wavelet analyses, which often employ hierarchical decompositions. Inspired by such principles, particularly the idea of signal separation, we introduce a diffusion framework leveraging multi-scale latent factorization. Our framework uniquely decomposes the denoising target, typically latent features from a pretrained Variational Autoencoder, into a low-frequency base signal capturing core structural information and a high-frequency residual signal that contributes finer, high-frequency details like textures. This decomposition into base and residual components directly informs our two-stage image generation process, which first produces the low-resolution base, followed by the generation of the high-resolution residual. Our proposed architecture facilitates reduced sampling steps during the residual learning stage, owing to the inherent ease of modeling residual information, which confers advantages over conventional full-resolution generation techniques. This specific approach of decomposing the signal into a base and a residual, conceptually akin to how wavelet analysis can separate different frequency bands, yields a more streamlined and intuitive design distinct from generic hierarchical models. Our method, \name\ (Multi-Scale Factorization), demonstrates its effectiveness by achieving FID scores of 2.08 ($256\times256$) and 2.47 ($512\times512$) on class-conditional ImageNet benchmarks, outperforming the DiT baseline (2.27 and 3.04 respectively) while also delivering a $4\times$ speed-up with the same number of sampling steps.

Yunjie Tian, Lingxi Xie, Xiaopeng Zhang et al.

In this paper, we propose a self-supervised visual representation learning approach which involves both generative and discriminative proxies, where we focus on the former part by requiring the target network to recover the original image based on the mid-level features. Different from prior work that mostly focuses on pixel-level similarity between the original and generated images, we advocate for Semantic-aware Generation (SaGe) to facilitate richer semantics rather than details to be preserved in the generated image. The core idea of implementing SaGe is to use an evaluator, a deep network that is pre-trained without labels, for extracting semantic-aware features. SaGe complements the target network with view-specific features and thus alleviates the semantic degradation brought by intensive data augmentations. We execute SaGe on ImageNet-1K and evaluate the pre-trained models on five downstream tasks including nearest neighbor test, linear classification, and fine-scaled image recognition, demonstrating its ability to learn stronger visual representations.

Bowen Shi, Xiaopeng Zhang, Haohang Xu et al.

Collecting annotated data for semantic segmentation is time-consuming and hard to scale up. In this paper, we for the first time propose a unified framework, termed as Multi-Dataset Pretraining, to take full advantage of the fragmented annotations of different datasets. The highlight is that the annotations from different domains can be efficiently reused and consistently boost performance for each specific domain. This is achieved by first pretraining the network via the proposed pixel-to-prototype contrastive loss over multiple datasets regardless of their taxonomy labels, and followed by fine-tuning the pretrained model over specific dataset as usual. In order to better model the relationship among images and classes from different datasets, we extend the pixel level embeddings via cross dataset mixing and propose a pixel-to-class sparse coding strategy that explicitly models the pixel-class similarity over the manifold embedding space. In this way, we are able to increase intra-class compactness and inter-class separability, as well as considering inter-class similarity across different datasets for better transferability. Experiments conducted on several benchmarks demonstrate its superior performance. Notably, MDP consistently outperforms the pretrained models over ImageNet by a considerable margin, while only using less than 10% samples for pretraining.

Yuhang Zhang, Xiaopeng Zhang, Robert. C. Qiu et al.

Semi-supervised learning acts as an effective way to leverage massive unlabeled data. In this paper, we propose a novel training strategy, termed as Semi-supervised Contrastive Learning (SsCL), which combines the well-known contrastive loss in self-supervised learning with the cross entropy loss in semi-supervised learning, and jointly optimizes the two objectives in an end-to-end way. The highlight is that different from self-training based semi-supervised learning that conducts prediction and retraining over the same model weights, SsCL interchanges the predictions over the unlabeled data between the two branches, and thus formulates a co-calibration procedure, which we find is beneficial for better prediction and avoid being trapped in local minimum. Towards this goal, the contrastive loss branch models pairwise similarities among samples, using the nearest neighborhood generated from the cross entropy branch, and in turn calibrates the prediction distribution of the cross entropy branch with the contrastive similarity. We show that SsCL produces more discriminative representation and is beneficial to few shot learning. Notably, on ImageNet with ResNet50 as the backbone, SsCL achieves 60.2% and 72.1% top-1 accuracy with 1% and 10% labeled samples, respectively, which significantly outperforms the baseline, and is better than previous semi-supervised and self-supervised methods.

Haohang Xu, Xiaopeng Zhang, Hao Li et al.

Self-supervised learning based on instance discrimination has shown remarkable progress. In particular, contrastive learning, which regards each image as well as its augmentations as an individual class and tries to distinguish them from all other images, has been verified effective for representation learning. However, pushing away two images that are de facto similar is suboptimal for general representation. In this paper, we propose a hierarchical semantic alignment strategy via expanding the views generated by a single image to \textbf{Cross-samples and Multi-level} representation, and models the invariance to semantically similar images in a hierarchical way. This is achieved by extending the contrastive loss to allow for multiple positives per anchor, and explicitly pulling semantically similar images/patches together at different layers of the network. Our method, termed as CsMl, has the ability to integrate multi-level visual representations across samples in a robust way. CsMl is applicable to current contrastive learning based methods and consistently improves the performance. Notably, using the moco as an instantiation, CsMl achieves a \textbf{76.6\% }top-1 accuracy with linear evaluation using ResNet-50 as backbone, and \textbf{66.7\%} and \textbf{75.1\%} top-1 accuracy with only 1\% and 10\% labels, respectively. \textbf{All these numbers set the new state-of-the-art.}

Haohang Xu, Hongkai Xiong, Guo-Jun Qi

In this paper, we propose the $K$-Shot Contrastive Learning (KSCL) of visual features by applying multiple augmentations to investigate the sample variations within individual instances. It aims to combine the advantages of inter-instance discrimination by learning discriminative features to distinguish between different instances, as well as intra-instance variations by matching queries against the variants of augmented samples over instances. Particularly, for each instance, it constructs an instance subspace to model the configuration of how the significant factors of variations in $K$-shot augmentations can be combined to form the variants of augmentations. Given a query, the most relevant variant of instances is then retrieved by projecting the query onto their subspaces to predict the positive instance class. This generalizes the existing contrastive learning that can be viewed as a special one-shot case. An eigenvalue decomposition is performed to configure instance subspaces, and the embedding network can be trained end-to-end through the differentiable subspace configuration. Experiment results demonstrate the proposed $K$-shot contrastive learning achieves superior performances to the state-of-the-art unsupervised methods.

Haohang Xu, Hongkai Xiong, Guojun Qi

One of the most significant challenges facing a few-shot learning task is the generalizability of the (meta-)model from the base to the novel categories. Most of existing few-shot learning models attempt to address this challenge by either learning the meta-knowledge from multiple simulated tasks on the base categories, or resorting to data augmentation by applying various transformations to training examples. However, the supervised nature of model training in these approaches limits their ability of exploring variations across different categories, thus restricting their cross-category generalizability in modeling novel concepts. To this end, we present a novel regularization mechanism by learning the change of feature representations induced by a distribution of transformations without using the labels of data examples. We expect this regularizer could expand the semantic space of base categories to cover that of novel categories through the transformation of feature representations. It could minimize the risk of overfitting into base categories by inspecting the transformation-augmented variations at the encoded feature level. This results in the proposed FLAT (Few-shot Learning via Autoencoding Transformations) approach by autoencoding the applied transformations. The experiment results show the superior performances to the current state-of-the-art methods in literature.

Feng Lin, Haohang Xu, Houqiang Li et al.

Self-supervised learning by predicting transformations has demonstrated outstanding performances in both unsupervised and (semi-)supervised tasks. Among the state-of-the-art methods is the AutoEncoding Transformations (AET) by decoding transformations from the learned representations of original and transformed images. Both deterministic and probabilistic AETs rely on the Euclidean distance to measure the deviation of estimated transformations from their groundtruth counterparts. However, this assumption is questionable as a group of transformations often reside on a curved manifold rather staying in a flat Euclidean space. For this reason, we should use the geodesic to characterize how an image transform along the manifold of a transformation group, and adopt its length to measure the deviation between transformations. Particularly, we present to autoencode a Lie group of homography transformations PG(2) to learn image representations. For this, we make an estimate of the intractable Riemannian logarithm by projecting PG(2) to a subgroup of rotation transformations SO(3) that allows the closed-form expression of geodesic distances. Experiments demonstrate the proposed AETv2 model outperforms the previous version as well as the other state-of-the-art self-supervised models in multiple tasks.