arXiv subjects

Hui Zeng

Publications and source records attributed to Hui Zeng.

At least 55 records · Page 3Linked to original sources

arXiv2025

Everywhere Attack: Attacking Locally and Globally to Boost Targeted Transferability

Adversarial examples' (AE) transferability refers to the phenomenon that AEs crafted with one surrogate model can also fool other models. Notwithstanding remarkable progress in untargeted transferability, its targeted counterpart remains challenging. This paper proposes an everywhere scheme to boost targeted transferability. Our idea is to attack a victim image both globally and locally. We aim to optimize 'an army of targets' in every local image region instead of the previous works that optimize a high-confidence target in the image. Specifically, we split a victim image into non-overlap blocks and jointly mount a targeted attack on each block. Such a strategy mitigates transfer failures caused by attention inconsistency between surrogate and victim models and thus results in stronger transferability. Our approach is method-agnostic, which means it can be easily combined with existing transferable attacks for even higher transferability. Extensive experiments on ImageNet demonstrate that the proposed approach universally improves the state-of-the-art targeted attacks by a clear margin, e.g., the transferability of the widely adopted Logit attack can be improved by 28.8%-300%.We also evaluate the crafted AEs on a real-world platform: Google Cloud Vision. Results further support the superiority of the proposed method.

cs.CV↗

arXiv2024

Two Heads Are Better Than One: Averaging along Fine-Tuning to Improve Targeted Transferability

With much longer optimization time than that of untargeted attacks notwithstanding, the transferability of targeted attacks is still far from satisfactory. Recent studies reveal that fine-tuning an existing adversarial example (AE) in feature space can efficiently boost its targeted transferability. However, existing fine-tuning schemes only utilize the endpoint and ignore the valuable information in the fine-tuning trajectory. Noting that the vanilla fine-tuning trajectory tends to oscillate around the periphery of a flat region of the loss surface, we propose averaging over the fine-tuning trajectory to pull the crafted AE towards a more centered region. We compare the proposed method with existing fine-tuning schemes by integrating them with state-of-the-art targeted attacks in various attacking scenarios. Experimental results uphold the superiority of the proposed method in boosting targeted transferability. The code is available at github.com/zengh5/Avg_FT.

cs.CV↗

arXiv2024

The local thermodynamic instability from negative susceptibility in a holographic superfluid with nonlinear terms

The local thermodynamic stability from the charge susceptibility of a holographic superfluid model at finite superfluid velocity is studied in the probe limit. Previous studies show that beyond a finite value of the superfluid velocity, the superfluid phase transition in the grand canonical ensemble becomes first order. We further reveal that in the canonical ensemble, the superfluid phase transition is still second order, and the difference indicates a section with negative susceptibility which means local thermodynamic instability beyond this superfluid velocity. However, we also meet the ``cave of wind'' behavior at larger superfluid velocity which complicate the phase diagram. We further study the influence of the two nonlinear terms $λ|ψ|^4$ and $τ|ψ|^6$ with parameters $λ$ and $τ$ on the condensate curves, and set appropriate values of $λ$ and $τ$ to remove the "cave of wind" region in the canonical ensemble to get a more elegant phase diagram and better represent the region with such instability, which is possible to be used to realize spontaneous formation of vortexes and quantum turbulence.

hep-th↗

arXiv2024

NTIRE 2024 Restore Any Image Model (RAIM) in the Wild Challenge

In this paper, we review the NTIRE 2024 challenge on Restore Any Image Model (RAIM) in the Wild. The RAIM challenge constructed a benchmark for image restoration in the wild, including real-world images with/without reference ground truth in various scenarios from real applications. The participants were required to restore the real-captured images from complex and unknown degradation, where generative perceptual quality and fidelity are desired in the restoration result. The challenge consisted of two tasks. Task one employed real referenced data pairs, where quantitative evaluation is available. Task two used unpaired images, and a comprehensive user study was conducted. The challenge attracted more than 200 registrations, where 39 of them submitted results with more than 400 submissions. Top-ranked methods improved the state-of-the-art restoration performance and obtained unanimous recognition from all 18 judges. The proposed datasets are available at https://drive.google.com/file/d/1DqbxUoiUqkAIkExu3jZAqoElr_nu1IXb/view?usp=sharing and the homepage of this challenge is at https://codalab.lisn.upsaclay.fr/competitions/17632.

cs.CV↗

arXiv2024

Intrinsic Orbital Origin for the Chirality-Dependent Nonlinear Planar Hall Effect of Topological Nodal Fermions in Chiral Crystals

Topological semimetals in chiral crystals, which possess both structural handedness and band crossings (or nodes) with topological chiral charge, exhibit many exotic physical properties. Here we demonstrate that the structural and electronic chirality of these systems can endow them with another fascinating phenomenon -- the intrinsic nonlinear planar Hall effect (INPHE), which is prominent around the nodes and reverses sign upon chirality reversal in opposite enantiomers. Taking chiral tellurium as an example, we reveal an intrinsic orbital mechanism, which manifests diverging orbital magnetic moments with hedgehog-like textures around nodes and, therefore, generates a dominant contribution to the INPHE that is proportional to the topological charge. Furthermore, we show that multifold fermions in topological chiral semimetals with B20 structures (e.g., CoSi and PtAl) induce a giant INPHE conductivity reaching the order of $1\sim 10\; \mathrm{A}\cdot\mathrm{V}^{-2}\cdot\mathrm{T}^{-1}$, which is detectable in experiments. Our study not only relates nonlinear transport to band topology and enantiomer recognition but also offers a new way to explore the exotic physical properties associated with unconventional chiral fermions.

cond-mat.mes-hall↗

arXiv2024

Enhancing targeted transferability via feature space fine-tuning

Adversarial examples (AEs) have been extensively studied due to their potential for privacy protection and inspiring robust neural networks. Yet, making a targeted AE transferable across unknown models remains challenging. In this paper, to alleviate the overfitting dilemma common in an AE crafted by existing simple iterative attacks, we propose fine-tuning it in the feature space. Specifically, starting with an AE generated by a baseline attack, we encourage the features conducive to the target class and discourage the features to the original class in a middle layer of the source model. Extensive experiments demonstrate that only a few iterations of fine-tuning can boost existing attacks' targeted transferability nontrivially and universally. Our results also verify that the simple iterative attacks can yield comparable or even better transferability than the resource-intensive methods, which rest on training target-specific classifiers or generators with additional data. The code is available at: github.com/zengh5/TA_feature_FT.

cs.CV↗

arXiv2023

Guided Image Restoration via Simultaneous Feature and Image Guided Fusion

Guided image restoration (GIR), such as guided depth map super-resolution and pan-sharpening, aims to enhance a target image using guidance information from another image of the same scene. Currently, joint image filtering-inspired deep learning-based methods represent the state-of-the-art for GIR tasks. Those methods either deal with GIR in an end-to-end way by elaborately designing filtering-oriented deep neural network (DNN) modules, focusing on the feature-level fusion of inputs; or explicitly making use of the traditional joint filtering mechanism by parameterizing filtering coefficients with DNNs, working on image-level fusion. The former ones are good at recovering contextual information but tend to lose fine-grained details, while the latter ones can better retain textual information but might lead to content distortions. In this work, to inherit the advantages of both methodologies while mitigating their limitations, we proposed a Simultaneous Feature and Image Guided Fusion (SFIGF) network, that simultaneously considers feature and image-level guided fusion following the guided filter (GF) mechanism. In the feature domain, we connect the cross-attention (CA) with GF, and propose a GF-inspired CA module for better feature-level fusion; in the image domain, we fully explore the GF mechanism and design GF-like structure for better image-level fusion. Since guided fusion is implemented in both feature and image domains, the proposed SFIGF is expected to faithfully reconstruct both contextual and textual information from sources and thus lead to better GIR results. We apply SFIGF to 4 typical GIR tasks, and experimental results on these tasks demonstrate its effectiveness and general availability.

cs.CV↗

arXiv2023

Generalization of the Nested Wilson Loop Formalism in Topological Dirac Semimetals with Higher-order Fermi Arcs

We generalize the nested Wilson loop formalism, which has been playing an important role in the study of topological quadrupole insulators, to two-dimensional (2D) and 3D nonsymmorphic materials with higher-order topology. In particular, certain 3D Dirac semimetals exhibit 1D higher-order Fermi arc (HOFA) states localizing on hinges where two surfaces meet and connecting the projection of the bulk Dirac points. We discover that the generalized nested Berry phase (gNBP) derived from this formalism is the bulk topological indicator determining the existence/absence of HOFAs, revealing a direct bulk-hinge correspondence in 3D Dirac semimetals. Finally, we study the Dirac semimetals NaCuSe and KMgBi based on first-principles calculations and explicitly show that the change in gNBP adjacent to the Dirac point corresponds to the termination of HOFAs at the projection of Dirac points on the hinge. Our findings not only improve the understanding of the bulk-hinge correspondence in topological Dirac semimetals but also provide a general formalism for studying the higher-order topology in nonsymmorphic systems.

cond-mat.mes-hall↗

arXiv2023

The holographic s+p model in 4D and 5D Einstein-Gauss-Bonnet gravity

We study the holographic s+p model in both four dimensional (4D) and five dimensional (5D) Einstein-Gauss-Bonnet (EGB) gravity. We first show a phase diagram with the Gauss-Bonnet parameter fixed to a small value $α=10^{-7}$ to choose propitiate values of $q_p/q_s$. Then we fix the value of $q_p/q_s$ and plot $α-μ$ phase diagrams to show the influence of Gauss-Bonnet term on the phase transitions in both 4D and 5D bulk, respectively. The phase diagrams in 4D and 5D present the same qualitative features, indicating similarity of 4D Einstein-Gauss-Bonnet gravity with the 5D case in holography. We also study the influences of Gauss-Bonnet parameter on the special values of the fourth order nonlinear term parameters $λ_s$ and $λ_p$, below which the condensate grows to a different direction near the critical point, that is important in realizing 1st order superfluid phase transitions. Especially, we notice that these special values are different in the canonical and grand canonical ensembles, which is closely related to the study of the spinodal region, where the phase separations occurs with the linear instability at finite wave vector.

hep-th↗

arXiv2023

Intrinsic Nonlinear Hall Detection of the Néel Vector for Two-Dimensional Antiferromagnetic Spintronics

The respective unique merit of antiferromagnets and two-dimensional (2D) materials in spintronic applications inspire us to exploit 2D antiferromagnetic spintronics. However, the detection of the Néel vector in 2D antiferromagnets remains a great challenge because the measured signals usually decrease significantly in the 2D limit. Here we propose that the Néel vector of 2D antiferromagnets can be efficiently detected by the intrinsic nonlinear Hall (INH) effect which exhibits unexpected significant signals. As a specific example, we show that the INH conductivity of the monolayer manganese chalcogenides Mn$X$ ($X$=S, Se, Te) can reach the order of nm$\cdot$mA/V$^2$, which is orders of magnitude larger than experimental values of paradigmatic antiferromagnetic spintronic materials. The INH effect can be accurately controlled by shifting the chemical potential around the band edge, which is experimentally feasible via electric gating or charge doping. Moreover, we explicitly demonstrate its $2π$-periodic dependence on the Néel vector orientation based on an effective $k.p$ model. Our findings enable flexible design schemes and promising material platforms for spintronic memory device applications based on 2D antiferromagnets.

cond-mat.mtrl-sci↗

arXiv2023

Human Guided Ground-truth Generation for Realistic Image Super-resolution

How to generate the ground-truth (GT) image is a critical issue for training realistic image super-resolution (Real-ISR) models. Existing methods mostly take a set of high-resolution (HR) images as GTs and apply various degradations to simulate their low-resolution (LR) counterparts. Though great progress has been achieved, such an LR-HR pair generation scheme has several limitations. First, the perceptual quality of HR images may not be high enough, limiting the quality of Real-ISR outputs. Second, existing schemes do not consider much human perception in GT generation, and the trained models tend to produce over-smoothed results or unpleasant artifacts. With the above considerations, we propose a human guided GT generation scheme. We first elaborately train multiple image enhancement models to improve the perceptual quality of HR images, and enable one LR image having multiple HR counterparts. Human subjects are then involved to annotate the high quality regions among the enhanced HR images as GTs, and label the regions with unpleasant artifacts as negative samples. A human guided GT image dataset with both positive and negative samples is then constructed, and a loss function is proposed to train the Real-ISR models. Experiments show that the Real-ISR models trained on our dataset can produce perceptually more realistic results with less artifacts. Dataset and codes can be found at https://github.com/ChrisDud0257/HGGT

cs.CV↗

arXiv2022

Efficient and Degradation-Adaptive Network for Real-World Image Super-Resolution

Efficient and effective real-world image super-resolution (Real-ISR) is a challenging task due to the unknown complex degradation of real-world images and the limited computation resources in practical applications. Recent research on Real-ISR has achieved significant progress by modeling the image degradation space; however, these methods largely rely on heavy backbone networks and they are inflexible to handle images of different degradation levels. In this paper, we propose an efficient and effective degradation-adaptive super-resolution (DASR) network, whose parameters are adaptively specified by estimating the degradation of each input image. Specifically, a tiny regression network is employed to predict the degradation parameters of the input image, while several convolutional experts with the same topology are jointly optimized to specify the network parameters via a non-linear mixture of experts. The joint optimization of multiple experts and the degradation-adaptive pipeline significantly extend the model capacity to handle degradations of various levels, while the inference remains efficient since only one adaptively specified network is used for super-resolving the input image. Our extensive experiments demonstrate that the proposed DASR is not only much more effective than existing methods on handling real-world images with different degradation levels but also efficient for easy deployment. Codes, models and datasets are available at https://github.com/csjliang/DASR.

cs.CV↗

arXiv2022

Details or Artifacts: A Locally Discriminative Learning Approach to Realistic Image Super-Resolution

Single image super-resolution (SISR) with generative adversarial networks (GAN) has recently attracted increasing attention due to its potentials to generate rich details. However, the training of GAN is unstable, and it often introduces many perceptually unpleasant artifacts along with the generated details. In this paper, we demonstrate that it is possible to train a GAN-based SISR model which can stably generate perceptually realistic details while inhibiting visual artifacts. Based on the observation that the local statistics (e.g., residual variance) of artifact areas are often different from the areas of perceptually friendly details, we develop a framework to discriminate between GAN-generated artifacts and realistic details, and consequently generate an artifact map to regularize and stabilize the model training process. Our proposed locally discriminative learning (LDL) method is simple yet effective, which can be easily plugged in off-the-shelf SISR methods and boost their performance. Experiments demonstrate that LDL outperforms the state-of-the-art GAN based SISR methods, achieving not only higher reconstruction accuracy but also superior perceptual quality on both synthetic and real-world datasets. Codes and models are available at https://github.com/csjliang/LDL.

eess.IV↗

arXiv2022

Efficient Long-Range Attention Network for Image Super-resolution

Recently, transformer-based methods have demonstrated impressive results in various vision tasks, including image super-resolution (SR), by exploiting the self-attention (SA) for feature extraction. However, the computation of SA in most existing transformer based models is very expensive, while some employed operations may be redundant for the SR task. This limits the range of SA computation and consequently the SR performance. In this work, we propose an efficient long-range attention network (ELAN) for image SR. Specifically, we first employ shift convolution (shift-conv) to effectively extract the image local structural information while maintaining the same level of complexity as 1x1 convolution, then propose a group-wise multi-scale self-attention (GMSA) module, which calculates SA on non-overlapped groups of features using different window sizes to exploit the long-range image dependency. A highly efficient long-range attention block (ELAB) is then built by simply cascading two shift-conv with a GMSA module, which is further accelerated by using a shared attention mechanism. Without bells and whistles, our ELAN follows a fairly simple design by sequentially cascading the ELABs. Extensive experiments demonstrate that ELAN obtains even better results against the transformer-based SR models but with significantly less complexity. The source code can be found at https://github.com/xindongzhang/ELAN.

cs.CV↗

arXiv2021

A comparison study of CNN denoisers on PRNU extraction

Performance of the sensor-based camera identification (SCI) method heavily relies on the denoising filter in estimating Photo-Response Non-Uniformity (PRNU). Given various attempts on enhancing the quality of the extracted PRNU, it still suffers from unsatisfactory performance in low-resolution images and high computational demand. Leveraging the similarity of PRNU estimation and image denoising, we take advantage of the latest achievements of Convolutional Neural Network (CNN)-based denoisers for PRNU extraction. In this paper, a comparative evaluation of such CNN denoisers on SCI performance is carried out on the public "Dresden Image Database". Our findings are two-fold. From one aspect, both the PRNU extraction and image denoising separate noise from the image content. Hence, SCI can benefit from the recent CNN denoisers if carefully trained. From another aspect, the goals and the scenarios of PRNU extraction and image denoising are different since one optimizes the quality of noise and the other optimizes the image quality. A carefully tailored training is needed when CNN denoisers are used for PRNU estimation. Alternative strategies of training data preparation and loss function design are analyzed theoretically and evaluated experimentally. We point out that feeding the CNNs with image-PRNU pairs and training them with correlation-based loss function result in the best PRNU estimation performance. To facilitate further studies of SCI, we also propose a minimum-loss camera fingerprint quantization scheme using which we save the fingerprints as image files in PNG format. Furthermore, we make the quantized fingerprints of the cameras from the "Dresden Image Database" publicly available.

eess.IV↗

arXiv2021

Holographic multi-condensate with nonlinear terms

We study the influence of nonlinear terms quartic of the charged fields, which do not change the critical points of single condensate solutions, on the phase structure of a holographic model with multi-condensate in probe limit. We include one s-wave order and one p-wave order charged under the same U(1) gauge field in the holographic model and study the influence of the three quartic nonlinear terms of the charged fields with coefficients $λ_s$, $λ_p$ and $λ_{sp}$ on the phase structure. We show the influence of each of the three parameters on the phase diagram with other two set to zero, respectively. With these nonlinear terms, we get more power on tuning the phase structure of the holographic system showing multi-condensate, and show how to get a reentrant phase transition as an example.

hep-th↗

arXiv2021

PPR10K: A Large-Scale Portrait Photo Retouching Dataset with Human-Region Mask and Group-Level Consistency

Different from general photo retouching tasks, portrait photo retouching (PPR), which aims to enhance the visual quality of a collection of flat-looking portrait photos, has its special and practical requirements such as human-region priority (HRP) and group-level consistency (GLC). HRP requires that more attention should be paid to human regions, while GLC requires that a group of portrait photos should be retouched to a consistent tone. Models trained on existing general photo retouching datasets, however, can hardly meet these requirements of PPR. To facilitate the research on this high-frequency task, we construct a large-scale PPR dataset, namely PPR10K, which is the first of its kind to our best knowledge. PPR10K contains $1, 681$ groups and $11, 161$ high-quality raw portrait photos in total. High-resolution segmentation masks of human regions are provided. Each raw photo is retouched by three experts, while they elaborately adjust each group of photos to have consistent tones. We define a set of objective measures to evaluate the performance of PPR and propose strategies to learn PPR models with good HRP and GLC performance. The constructed PPR10K dataset provides a good benchmark for studying automatic PPR methods, and experiments demonstrate that the proposed learning strategies are effective to improve the retouching performance. Datasets and codes are available: https://github.com/csjliang/PPR10K.

cs.CV↗

arXiv2021

High-Resolution Photorealistic Image Translation in Real-Time: A Laplacian Pyramid Translation Network

Existing image-to-image translation (I2IT) methods are either constrained to low-resolution images or long inference time due to their heavy computational burden on the convolution of high-resolution feature maps. In this paper, we focus on speeding-up the high-resolution photorealistic I2IT tasks based on closed-form Laplacian pyramid decomposition and reconstruction. Specifically, we reveal that the attribute transformations, such as illumination and color manipulation, relate more to the low-frequency component, while the content details can be adaptively refined on high-frequency components. We consequently propose a Laplacian Pyramid Translation Network (LPTN) to simultaneously perform these two tasks, where we design a lightweight network for translating the low-frequency component with reduced resolution and a progressive masking strategy to efficiently refine the high-frequency ones. Our model avoids most of the heavy computation consumed by processing high-resolution feature maps and faithfully preserves the image details. Extensive experimental results on various tasks demonstrate that the proposed method can translate 4K images in real-time using one normal GPU while achieving comparable transformation performance against existing methods. Datasets and codes are available: https://github.com/csjliang/LPTN.

cs.CV↗