Search arXiv⌕ Search

arXiv subjects

Chang Xu

Publications and source records attributed to Chang Xu.

At least 253 records · Page 14Linked to original sources
arXiv2023

Dynamic Coarse-to-Fine Learning for Oriented Tiny Object Detection

Detecting arbitrarily oriented tiny objects poses intense challenges to existing detectors, especially for label assignment. Despite the exploration of adaptive label assignment in recent oriented object detectors, the extreme geometry shape and limited feature of oriented tiny objects still induce severe mismatch and imbalance issues. Specifically, the position prior, positive sample feature, and instance are mismatched, and the learning of extreme-shaped objects is biased and unbalanced due to little proper feature supervision. To tackle these issues, we propose a dynamic prior along with the coarse-to-fine assigner, dubbed DCFL. For one thing, we model the prior, label assignment, and object representation all in a dynamic manner to alleviate the mismatch issue. For another, we leverage the coarse prior matching and finer posterior constraint to dynamically assign labels, providing appropriate and relatively balanced supervision for diverse instances. Extensive experiments on six datasets show substantial improvements to the baseline. Notably, we obtain the state-of-the-art performance for one-stage detectors on the DOTA-v1.5, DOTA-v2.0, and DIOR-R datasets under single-scale training and testing. Codes are available at https://github.com/Chasel-Tsui/mmrotate-dcfl.

cs.CV↗
arXiv2023

PSVT: End-to-End Multi-person 3D Pose and Shape Estimation with Progressive Video Transformers

Existing methods of multi-person video 3D human Pose and Shape Estimation (PSE) typically adopt a two-stage strategy, which first detects human instances in each frame and then performs single-person PSE with temporal model. However, the global spatio-temporal context among spatial instances can not be captured. In this paper, we propose a new end-to-end multi-person 3D Pose and Shape estimation framework with progressive Video Transformer, termed PSVT. In PSVT, a spatio-temporal encoder (STE) captures the global feature dependencies among spatial objects. Then, spatio-temporal pose decoder (STPD) and shape decoder (STSD) capture the global dependencies between pose queries and feature tokens, shape queries and feature tokens, respectively. To handle the variances of objects as time proceeds, a novel scheme of progressive decoding is used to update pose and shape queries at each frame. Besides, we propose a novel pose-guided attention (PGA) for shape decoder to better predict shape parameters. The two components strengthen the decoder of PSVT to improve performance. Extensive experiments on the four datasets show that PSVT achieves stage-of-the-art results.

cs.CV↗
arXiv2023

Masked Distillation with Receptive Tokens

Distilling from the feature maps can be fairly effective for dense prediction tasks since both the feature discriminability and localization priors can be well transferred. However, not every pixel contributes equally to the performance, and a good student should learn from what really matters to the teacher. In this paper, we introduce a learnable embedding dubbed receptive token to localize those pixels of interests (PoIs) in the feature map, with a distillation mask generated via pixel-wise attention. Then the distillation will be performed on the mask via pixel-wise reconstruction. In this way, a distillation mask actually indicates a pattern of pixel dependencies within feature maps of teacher. We thus adopt multiple receptive tokens to investigate more sophisticated and informative pixel dependencies to further enhance the distillation. To obtain a group of masks, the receptive tokens are learned via the regular task loss but with teacher fixed, and we also leverage a Dice loss to enrich the diversity of learned masks. Our method dubbed MasKD is simple and practical, and needs no priors of tasks in application. Experiments show that our MasKD can achieve state-of-the-art performance consistently on object detection and semantic segmentation benchmarks. Code is available at: https://github.com/hunto/MasKD .

cs.CV↗
arXiv2023

Thin Films on the Skin, but not Frictional Agents, Attenuate the Percept of Pleasantness to Brushed Stimuli

Brushed stimuli are perceived as pleasant when stroked lightly on the skin surface of a touch receiver at certain velocities. While the relationship between brush velocity and pleasantness has been widely replicated, we do not understand how resultant skin movements - e.g., lateral stretch, stick-slip, normal indentation - drive us to form such judgments. In a series of psychophysical experiments, this work modulates skin movements by varying stimulus stiffness and employing various treatments. The stimuli include brushes of three levels of stiffness and an ungloved human finger. The skin's friction is modulated via non-hazardous chemicals and washing protocols, and the skin's thickness and lateral movement are modulated by thin sheets of adhesive film. The stimuli are hand-brushed at controlled forces and velocities. Human participants report perceived pleasantness per trial using ratio scaling. The results indicate that a brush's stiffness influenced pleasantness more than any skin treatment. Surprisingly, varying the skin's friction did not affect pleasantness. However, the application of a thin elastic film modulated pleasantness. Such barriers, though elastic and only 40 microns thick, inhibit the skin's tangential movement and disperse normal force. The finding that thin films modulate affective interactions has implications for wearable sensors and actuation devices.

q-bio.NC↗
arXiv2023

Two-in-one Knowledge Distillation for Efficient Facial Forgery Detection

Facial forgery detection is a crucial but extremely challenging topic, with the fast development of forgery techniques making the synthetic artefact highly indistinguishable. Prior works show that by mining both spatial and frequency information the forgery detection performance of deep learning models can be vastly improved. However, leveraging multiple types of information usually requires more than one branch in the neural network, which makes the model heavy and cumbersome. Knowledge distillation, as an important technique for efficient modelling, could be a possible remedy. We find that existing knowledge distillation methods have difficulties distilling a dual-branch model into a single-branch model. More specifically, knowledge distillation on both the spatial and frequency branches has degraded performance than distillation only on the spatial branch. To handle such problem, we propose a novel two-in-one knowledge distillation framework which can smoothly merge the information from a large dual-branch network into a small single-branch network, with the help of different dedicated feature projectors and the gradient homogenization technique. Experimental analysis on two datasets, FaceForensics++ and Celeb-DF, shows that our proposed framework achieves superior performance for facial forgery detection with much fewer parameters.

cs.CV↗
arXiv2023

Anti-Compression Contrastive Facial Forgery Detection

Forgery facial images and videos have increased the concern of digital security. It leads to the significant development of detecting forgery data recently. However, the data, especially the videos published on the Internet, are usually compressed with lossy compression algorithms such as H.264. The compressed data could significantly degrade the performance of recent detection algorithms. The existing anti-compression algorithms focus on enhancing the performance in detecting heavily compressed data but less consider the compression adaption to the data from various compression levels. We believe creating a forgery detection model that can handle the data compressed with unknown levels is important. To enhance the performance for such models, we consider the weak compressed and strong compressed data as two views of the original data and they should have similar representation and relationships with other samples. We propose a novel anti-compression forgery detection framework by maintaining closer relations within data under different compression levels. Specifically, the algorithm measures the pair-wise similarity within data as the relations, and forcing the relations of weak and strong compressed data close to each other, thus improving the discriminate power for detecting strong compressed data. To achieve a better strong compressed data relation guided by the less compressed one, we apply video level contrastive learning for weak compressed data, which forces the model to produce similar representations within the same video and far from the negative samples. The experiment results show that the proposed algorithm could boost performance for strong compressed data while improving the accuracy rate when detecting the clean data.

cs.CV↗
arXiv2023

Unlabeled Imperfect Demonstrations in Adversarial Imitation Learning

Adversarial imitation learning has become a widely used imitation learning framework. The discriminator is often trained by taking expert demonstrations and policy trajectories as examples respectively from two categories (positive vs. negative) and the policy is then expected to produce trajectories that are indistinguishable from the expert demonstrations. But in the real world, the collected expert demonstrations are more likely to be imperfect, where only an unknown fraction of the demonstrations are optimal. Instead of treating imperfect expert demonstrations as absolutely positive or negative, we investigate unlabeled imperfect expert demonstrations as they are. A positive-unlabeled adversarial imitation learning algorithm is developed to dynamically sample expert demonstrations that can well match the trajectories from the constantly optimized agent policy. The trajectories of an initial agent policy could be closer to those non-optimal expert demonstrations, but within the framework of adversarial imitation learning, agent policy will be optimized to cheat the discriminator and produce trajectories that are similar to those optimal expert demonstrations. Theoretical analysis shows that our method learns from the imperfect demonstrations via a self-paced way. Experimental results on MuJoCo and RoboSuite platforms demonstrate the effectiveness of our method from different aspects.

cs.LG↗
arXiv2023

Adversarially Robust Neural Architectures

Deep Neural Networks (DNNs) are vulnerable to adversarial attacks. Existing methods are devoted to developing various robust training strategies or regularizations to update the weights of the neural network. But beyond the weights, the overall structure and information flow in the network are explicitly determined by the neural architecture, which remains unexplored. This paper thus aims to improve the adversarial robustness of the network from the architecture perspective. We explore the relationship among adversarial robustness, Lipschitz constant, and architecture parameters and show that an appropriate constraint on architecture parameters could reduce the Lipschitz constant to further improve the robustness. The importance of architecture parameters could vary from operation to operation or connection to connection. We approximate the Lipschitz constant of the entire network through a univariate log-normal distribution, whose mean and variance are related to architecture parameters. The confidence can be fulfilled through formulating a constraint on the distribution parameters based on the cumulative function. Compared with adversarially trained neural architectures searched by various NAS algorithms as well as efficient human-designed models, our algorithm empirically achieves the best performance among all the models under various attacks on different datasets.

cs.CV↗
arXiv2022

Knowledge Distillation from A Stronger Teacher

Unlike existing knowledge distillation methods focus on the baseline settings, where the teacher models and training strategies are not that strong and competing as state-of-the-art approaches, this paper presents a method dubbed DIST to distill better from a stronger teacher. We empirically find that the discrepancy of predictions between the student and a stronger teacher may tend to be fairly severer. As a result, the exact match of predictions in KL divergence would disturb the training and make existing methods perform poorly. In this paper, we show that simply preserving the relations between the predictions of teacher and student would suffice, and propose a correlation-based loss to capture the intrinsic inter-class relations from the teacher explicitly. Besides, considering that different instances have different semantic similarities to each class, we also extend this relational match to the intra-class level. Our method is simple yet practical, and extensive experiments demonstrate that it adapts well to various architectures, model sizes and training strategies, and can achieve state-of-the-art performance consistently on image classification, object detection, and semantic segmentation tasks. Code is available at: https://github.com/hunto/DIST_KD .

cs.CV↗
arXiv2022

Learning Spatiotemporal Frequency-Transformer for Low-Quality Video Super-Resolution

Video Super-Resolution (VSR) aims to restore high-resolution (HR) videos from low-resolution (LR) videos. Existing VSR techniques usually recover HR frames by extracting pertinent textures from nearby frames with known degradation processes. Despite significant progress, grand challenges are remained to effectively extract and transmit high-quality textures from high-degraded low-quality sequences, such as blur, additive noises, and compression artifacts. In this work, a novel Frequency-Transformer (FTVSR) is proposed for handling low-quality videos that carry out self-attention in a combined space-time-frequency domain. First, video frames are split into patches and each patch is transformed into spectral maps in which each channel represents a frequency band. It permits a fine-grained self-attention on each frequency band, so that real visual texture can be distinguished from artifacts. Second, a novel dual frequency attention (DFA) mechanism is proposed to capture the global frequency relations and local frequency relations, which can handle different complicated degradation processes in real-world scenarios. Third, we explore different self-attention schemes for video processing in the frequency domain and discover that a ``divided attention'' which conducts a joint space-frequency attention before applying temporal-frequency attention, leads to the best video enhancement quality. Extensive experiments on three widely-used VSR datasets show that FTVSR outperforms state-of-the-art methods on different low-quality videos with clear visual margins. Code and pre-trained models are available at https://github.com/researchmm/FTVSR.

eess.IV↗
arXiv2022

Boosting Semi-Supervised Semantic Segmentation with Probabilistic Representations

Recent breakthroughs in semi-supervised semantic segmentation have been developed through contrastive learning. In prevalent pixel-wise contrastive learning solutions, the model maps pixels to deterministic representations and regularizes them in the latent space. However, there exist inaccurate pseudo-labels which map the ambiguous representations of pixels to the wrong classes due to the limited cognitive ability of the model. In this paper, we define pixel-wise representations from a new perspective of probability theory and propose a Probabilistic Representation Contrastive Learning (PRCL) framework that improves representation quality by taking its probability into consideration. Through modelling the mapping from pixels to representations as the probability via multivariate Gaussian distributions, we can tune the contribution of the ambiguous representations to tolerate the risk of inaccurate pseudo-labels. Furthermore, we define prototypes in the form of distributions, which indicates the confidence of a class, while the point prototype cannot. Moreover, we propose to regularize the distribution variance to enhance the reliability of representations. Taking advantage of these benefits, high-quality feature representations can be derived in the latent space, thereby the performance of semantic segmentation can be further improved. We conduct sufficient experiment to evaluate PRCL on Pascal VOC and CityScapes to demonstrate its superiority. The code is available at https://github.com/Haoyu-Xie/PRCL.

cs.CV↗
arXiv2022

ContraFeat: Contrasting Deep Features for Semantic Discovery

StyleGAN has shown strong potential for disentangled semantic control, thanks to its special design of multi-layer intermediate latent variables. However, existing semantic discovery methods on StyleGAN rely on manual selection of modified latent layers to obtain satisfactory manipulation results, which is tedious and demanding. In this paper, we propose a model that automates this process and achieves state-of-the-art semantic discovery performance. The model consists of an attention-equipped navigator module and losses contrasting deep-feature changes. We propose two model variants, with one contrasting samples in a binary manner, and another one contrasting samples with learned prototype variation patterns. The proposed losses are defined with pretrained deep features, based on our assumption that the features can implicitly reveal the desired semantic structure including consistency and orthogonality. Additionally, we design two metrics to quantitatively evaluate the performance of semantic discovery methods on FFHQ dataset, and also show that disentangled representations can be derived via a simple training process. Experimentally, our models can obtain state-of-the-art semantic discovery results without relying on latent layer-wise manual selection, and these discovered semantics can be used to manipulate real-world images.

cs.CV↗
arXiv2022

FastMIM: Expediting Masked Image Modeling Pre-training for Vision

The combination of transformers and masked image modeling (MIM) pre-training framework has shown great potential in various vision tasks. However, the pre-training computational budget is too heavy and withholds the MIM from becoming a practical training paradigm. This paper presents FastMIM, a simple and generic framework for expediting masked image modeling with the following two steps: (i) pre-training vision backbones with low-resolution input images; and (ii) reconstructing Histograms of Oriented Gradients (HOG) feature instead of original RGB values of the input images. In addition, we propose FastMIM-P to progressively enlarge the input resolution during pre-training stage to further enhance the transfer results of models with high capacity. We point out that: (i) a wide range of input resolutions in pre-training phase can lead to similar performances in fine-tuning phase and downstream tasks such as detection and segmentation; (ii) the shallow layers of encoder are more important during pre-training and discarding last several layers can speed up the training stage with no harm to fine-tuning performance; (iii) the decoder should match the size of selected network; and (iv) HOG is more stable than RGB values when resolution transfers;. Equipped with FastMIM, all kinds of vision backbones can be pre-trained in an efficient way. For example, we can achieve 83.8%/84.1% top-1 accuracy on ImageNet-1K with ViT-B/Swin-B as backbones. Compared to previous relevant approaches, we can achieve comparable or better top-1 accuracy while accelerate the training procedure by $\sim$5$\times$. Code can be found in https://github.com/ggjy/FastMIM.pytorch.

cs.CV↗
arXiv2022

Non-interactive Multi-client Searchable Symmetric Encryption with Small Client Storage

Considerable attention has been paid to dynamic searchable symmetric encryption (DSSE) which allows users to search on dynamically updated encrypted databases. To improve the performance of real-world applications, recent non-interactive multi-client DSSE schemes are targeted at avoiding per-query interaction between data owners and data users. However, existing non-interactive multi-client DSSE schemes do not consider forward privacy or backward privacy, making them exposed to leakage abuse attacks. Besides, most existing DSSE schemes with forward and backward privacy rely on keeping a keyword operation counter or an inverted index, resulting in a heavy storage burden on the data owner side. To address these issues, we propose a non-interactive multi-client DSSE scheme with small client storage, and our proposed scheme can provide both forward privacy and backward privacy. Specifically, we first design a lightweight storage chain structure that binds all keywords to a single state to reduce the storage cost. Then, we present a Hidden Key technique, which preserves non-interactive forward privacy through time range queries, ensuring that data with newer timestamps cannot match earlier time ranges. We conduct extensive experiments to validate our methods, which demonstrate computational efficiency. Moreover, security analysis proves the privacy-preserving property of our methods.

cs.CR↗
arXiv2022

Competing topological superconducting phases in FeSe$_{0.45}$Te$_{0.55}$

We demonstrate that recent angle-resolved photo-emission spectroscopy experiments provide strong evidence for the existence of two competing topological superconducting phases in FeSe$_{0.45}$Te$_{0.55}$. The coupling of their underlying microscopic mechanisms -- one based on a three-dimensional topological insulator, one based on two-dimensional superconductivity -- stabilizes topological superconductivity over a wide range of parameters, and gives rise to two disjoint topological regions in the phase diagram of FeSe$_{0.45}$Te$_{0.55}$. We show that the topological origin of these regions can be identified by considering the form of Majorana edge modes at domain walls.

cond-mat.supr-con↗
arXiv2022

GhostNetV2: Enhance Cheap Operation with Long-Range Attention

Light-weight convolutional neural networks (CNNs) are specially designed for applications on mobile devices with faster inference speed. The convolutional operation can only capture local information in a window region, which prevents performance from being further improved. Introducing self-attention into convolution can capture global information well, but it will largely encumber the actual speed. In this paper, we propose a hardware-friendly attention mechanism (dubbed DFC attention) and then present a new GhostNetV2 architecture for mobile applications. The proposed DFC attention is constructed based on fully-connected layers, which can not only execute fast on common hardware but also capture the dependence between long-range pixels. We further revisit the expressiveness bottleneck in previous GhostNet and propose to enhance expanded features produced by cheap operations with DFC attention, so that a GhostNetV2 block can aggregate local and long-range information simultaneously. Extensive experiments demonstrate the superiority of GhostNetV2 over existing architectures. For example, it achieves 75.3% top-1 accuracy on ImageNet with 167M FLOPs, significantly suppressing GhostNetV1 (74.5%) with a similar computational cost. The source code will be available at https://github.com/huawei-noah/Efficient-AI-Backbones/tree/master/ghostnetv2_pytorch and https://gitee.com/mindspore/models/tree/master/research/cv/ghostnetv2.

cs.CV↗
arXiv2022

Effects of nucleon-nucleon short-range correlations on inclusive electron scattering

The nucleon-nucleon short-range correlation NN-SRC is one of the key issues of nuclear physics, which typically manifest themselves in high-momentum components of the nuclear momentum distributions. In this letter, the nuclear spectral functions based on the axially deformed relativistic mean-field model are developed to involve the NN-SRC. With the spectral functions, the inclusive electron scattering $ (e,e^{\prime}) $ cross sections are calculated within the PWIA framework, including the quasi-elastic (QE) part and $ Δ$ production part. Especially in the $ Δ$ production region, we reconsider the electromagnetic structures of the nucleon resonance $ Δ$(1232) and the scattering mechanisms, thereby the theoretical calculations are improved effectively and the cross sections are well consistent with the experimental data. The theoretical $ (e,e^{\prime}) $ cross sections are further divided into NN-SRC and mean-field contributions. It is found that, at the kinematics $ 0.5 \,{\rm GeV}^{2}<Q^{2}<1 \,{\rm GeV}^{2} $, the QE peak and $ Δ$ production peak not only reflect the mean-field structure, but also are sensitive to the NN-SRC information. Finally, we provide a new method to extract the strengths of NN-SRC from experimental cross sections for selected nuclei at the suitable kinematics.

nucl-th↗
arXiv2022

Exploring sensitivity of charge-exchange ($p, n$) reactions to the neutron density distribution

$Background:$ The determination of the nuclear neutron properties suffers from uncontrolled uncertainties, which attracted considerable attention recently, such as in the context of the PREX experiment. $Purpose:$ Our aim is to analyze the sensitivity of charge-exchange ($p, n$) reactions to the neutron density distribution $ρ_{n}$ and constrain the neutron characteristics in the nuclear structure models. $Method:$ By combing the folding and the mean-field models, the nucleon-nucleus ($NA$) potential can be obtained from the nuclear density distribution. Further, the ($p, p$) and ($p, n$) cross sections for $^{48}$Ca and $^{208}$Pb are calculated following the distorted-wave Born approximation (DWBA) method. $Results:$ Compared with the ($p, p$) cross section, the effects of $ρ_{n}$ variation on the ($p, n$) cross section are significant, which is due to the impact of isovector properties. Based on the global folding model analyses of data, it is found that $^{48}$Ca and $^{208}$Pb have relatively large neutron skin thickness $ΔR_{n p}$. $Conclusions:$ Results illustrate that the charge-exchange ($p, n$) reaction is a sensitive probe of $ρ_{n}$. The results in this paper can offer useful guides for future experiments of neutron characteristics.

nucl-th↗