Searcharxiv⌕ Search

arXiv subjects

Xin Liao

Publications and source records attributed to Xin Liao.

69 records · Page 4Linked to original sources
arXiv2023

SepMark: Deep Separable Watermarking for Unified Source Tracing and Deepfake Detection

Malicious Deepfakes have led to a sharp conflict over distinguishing between genuine and forged faces. Although many countermeasures have been developed to detect Deepfakes ex-post, undoubtedly, passive forensics has not considered any preventive measures for the pristine face before foreseeable manipulations. To complete this forensics ecosystem, we thus put forward the proactive solution dubbed SepMark, which provides a unified framework for source tracing and Deepfake detection. SepMark originates from encoder-decoder-based deep watermarking but with two separable decoders. For the first time the deep separable watermarking, SepMark brings a new paradigm to the established study of deep watermarking, where a single encoder embeds one watermark elegantly, while two decoders can extract the watermark separately at different levels of robustness. The robust decoder termed Tracer that resists various distortions may have an overly high level of robustness, allowing the watermark to survive both before and after Deepfake. The semi-robust one termed Detector is selectively sensitive to malicious distortions, making the watermark disappear after Deepfake. Only SepMark comprising of Tracer and Detector can reliably trace the trusted source of the marked face and detect whether it has been altered since being marked; neither of the two alone can achieve this. Extensive experiments demonstrate the effectiveness of the proposed SepMark on typical Deepfakes, including face swapping, expression reenactment, and attribute editing.

cs.CV↗
arXiv2023

Semi-supervised Pathological Image Segmentation via Cross Distillation of Multiple Attentions

Segmentation of pathological images is a crucial step for accurate cancer diagnosis. However, acquiring dense annotations of such images for training is labor-intensive and time-consuming. To address this issue, Semi-Supervised Learning (SSL) has the potential for reducing the annotation cost, but it is challenged by a large number of unlabeled training images. In this paper, we propose a novel SSL method based on Cross Distillation of Multiple Attentions (CDMA) to effectively leverage unlabeled images. Firstly, we propose a Multi-attention Tri-branch Network (MTNet) that consists of an encoder and a three-branch decoder, with each branch using a different attention mechanism that calibrates features in different aspects to generate diverse outputs. Secondly, we introduce Cross Decoder Knowledge Distillation (CDKD) between the three decoder branches, allowing them to learn from each other's soft labels to mitigate the negative impact of incorrect pseudo labels in training. Additionally, uncertainty minimization is applied to the average prediction of the three branches, which further regularizes predictions on unlabeled images and encourages inter-branch consistency. Our proposed CDMA was compared with eight state-of-the-art SSL methods on the public DigestPath dataset, and the experimental results showed that our method outperforms the other approaches under different annotation ratios. The code is available at \href{https://github.com/HiLab-git/CDMA}{https://github.com/HiLab-git/CDMA.}

cs.CV↗
arXiv2023

Exploiting Fine-Grained DCT Representations for Hiding Image-Level Messages within JPEG Images

Unlike hiding bit-level messages, hiding image-level messages is more challenging, which requires large capacity, high imperceptibility, and high security. Although recent advances in hiding image-level messages have been remarkable, existing schemes are limited to lossless spatial images as covers and cannot be directly applied to JPEG images, the ubiquitous lossy format images in daily life. The difficulties of migration are caused by the lack of targeted design and the loss of details due to lossy decompression and re-compression. Considering that taking DCT densely on $8\times8$ image patches is the core of the JPEG compression standard, we design a novel model called \textsf{EFDR}, which can comprehensively \underline{E}xploit \underline{F}ine-grained \underline{D}CT \underline{R}epresentations and embed the secret image into quantized DCT coefficients to avoid the lossy process. Specifically, we transform the JPEG cover image and hidden secret image into fine-grained DCT representations that compact the frequency and are associated with the inter-block and intra-block correlations. Subsequently, the fine-grained DCT representations are further enhanced by a sub-band features enhancement module. Afterward, a transformer-based invertibility module is designed to fuse enhanced sub-band features. Such a design enables a fine-grained self-attention on each sub-band and captures long-range dependencies while maintaining excellent reversibility for hiding and recovery. To our best knowledge, this is the first attempt to embed a color image of equal size in a color JPEG image. Extensive experiments demonstrate the effectiveness of our \textsf{EFDR} with superior performance.

cs.CV↗
arXiv2023

Mover: Mask and Recovery based Facial Part Consistency Aware Method for Deepfake Video Detection

Deepfake techniques have been widely used for malicious purposes, prompting extensive research interest in developing Deepfake detection methods. Deepfake manipulations typically involve tampering with facial parts, which can result in inconsistencies across different parts of the face. For instance, Deepfake techniques may change smiling lips to an upset lip, while the eyes remain smiling. Existing detection methods depend on specific indicators of forgery, which tend to disappear as the forgery patterns are improved. To address the limitation, we propose Mover, a new Deepfake detection model that exploits unspecific facial part inconsistencies, which are inevitable weaknesses of Deepfake videos. Mover randomly masks regions of interest (ROIs) and recovers faces to learn unspecific features, which makes it difficult for fake faces to be recovered, while real faces can be easily recovered. Specifically, given a real face image, we first pretrain a masked autoencoder to learn facial part consistency by dividing faces into three parts and randomly masking ROIs, which are then recovered based on the unmasked facial parts. Furthermore, to maximize the discrepancy between real and fake videos, we propose a novel model with dual networks that utilize the pretrained encoder and masked autoencoder, respectively. 1) The pretrained encoder is finetuned for capturing the encoding of inconsistent information in the given video. 2) The pretrained masked autoencoder is utilized for mapping faces and distinguishing real and fake videos. Our extensive experiments on standard benchmarks demonstrate that Mover is highly effective.

cs.MM↗
arXiv2023

Mover: Mask and Recovery based Facial Part Consistency Aware Method for Deepfake Video Detection

Deepfake techniques have been widely used for malicious purposes, prompting extensive research interest in developing Deepfake detection methods. Deepfake manipulations typically involve tampering with facial parts, which can result in inconsistencies across different parts of the face. For instance, Deepfake techniques may change smiling lips to an upset lip, while the eyes remain smiling. Existing detection methods depend on specific indicators of forgery, which tend to disappear as the forgery patterns are improved. To address the limitation, we propose Mover, a new Deepfake detection model that exploits unspecific facial part inconsistencies, which are inevitable weaknesses of Deepfake videos. Mover randomly masks regions of interest (ROIs) and recovers faces to learn unspecific features, which makes it difficult for fake faces to be recovered, while real faces can be easily recovered. Specifically, given a real face image, we first pretrain a masked autoencoder to learn facial part consistency by dividing faces into three parts and randomly masking ROIs, which are then recovered based on the unmasked facial parts. Furthermore, to maximize the discrepancy between real and fake videos, we propose a novel model with dual networks that utilize the pretrained encoder and masked autoencoder, respectively. 1) The pretrained encoder is finetuned for capturing the encoding of inconsistent information in the given video. 2) The pretrained masked autoencoder is utilized for mapping faces and distinguishing real and fake videos. Our extensive experiments on standard benchmarks demonstrate that Mover is highly effective.

cs.CV↗
arXiv2022

Manipulating single excess electrons in monolayer transition metal dihalide

Polarons are entities of excess electrons dressed with local response of lattices, whose atomic-scale characterization is essential for understanding the many body physics arising from the electron-lattice entanglement, but yet difficult to achieve. Here, using scanning tunneling microscopy and spectroscopy (STM/STS), we show the visualization and manipulation of single polarons with different origin, i.e., electronic and conventional polarons, in monolayer CoCl2, that are grown on HOPG substrate via molecular beam epitaxy. Four types of polarons are identified, all inducing upward local band bending, but exhibiting distinct appearances, lattice occupations, polaronic states and local lattice distortions. First principles calculations unveil three types of polarons are stabilized by electron-electron interaction. The type-4 polaron, however, are driven by conventional lattice distortions. All the four types of polarons can be created, moved, erased, and moreover interconverted individually by the STM tip, allowing precise control of single polarons unprecedently. This finding identifies the rich category of polarons and their feasibility of manipulation in CoCl2, which can be generalized to other transition metal halides.

cond-mat.mes-hall↗
arXiv2021

Mott insulator tuning via structural distortion in monolayer 1T-NbSe2

Mott state in 1T-TaS2 is predicted to host quantum spin liquids (QSL). However, its insulating mechanism is controversial due to complications from interlayer coupling. Here, we study the Mott state in monolayer 1T-NbSe2, an electronic analogy to TaS2 exempt from interlayer coupling, using spectroscopic imaging scanning tunneling microscopy and first principles calculations. Monolayer NbSe2 surprisingly displays two types of Star-of-David (SD) motifs with different Mott gap sizes, that are interconvertible via temperature variation. And, bilayer 1T-NbSe2 shows Mott collapse by interlayer coupling. Our calculation unveils the two types of SDs possess distinct structural distortions, altering the effective Coulomb energies of the central Nb orbital. Our calculation suggests the Mott gap, the same parameter for determining the QSL regime, is tunable with strain. This finding offers a general strategy for manipulating the Mott state in 1T-NbSe2 and related systems via structural distortions, which may be tuned into the potential QSL regime.

cond-mat.mes-hall↗
arXiv2017

Second-order expansions for maxima of dynamic bivariate normal copulas

In this paper, we establish the second-order distributional expansions of normalized maxima of n independent observations, where the ith observation follows from a normal copula with its correlation coefficient being a monotone continuous function. These expansions can be used to deduce the convergence rates of distributions of normalized maxima to their limits.

math.PR↗
arXiv2016

Second-order asymptotics on distributions of maxima of bivariate elliptical arrays

Let $\{ (ξ_{ni}, η_{ni}), 1\leq i \leq n, n\geq 1 \}$ be a triangular array of independent bivariate elliptical random vectors with the same distribution function as $(S_{1}, ρ_{n}S_{1}+\sqrt{1-ρ_{n}^2}S_{2})$, $ρ_{n}\in (0,1)$, where $(S_{1},S_{2})$ is a bivariate spherical random vector. For the distribution function of radius $\sqrt{S_{1}^2+S_{2}^2}$ belonging to the max-domain of attraction of the Weibull distribution, Hashorva (2006) derived the limiting distribution of maximum of this triangular array if convergence rate of $ρ_{n}$ to $1$ is given. In this paper, under the refinement of the rate of convergence of $ρ_{n}$ to $1$ and the second-order regular variation of the distributional tail of radius, precise second-order distributional expansions of the normalized maxima of bivariate elliptical triangular arrays are established.

math.PR↗
arXiv2016

Higher-order expansions of extremes from mixed skew-t distribution

In this paper, we study the asymptotic behaviors of the extreme of mixed skew-t distribution. We considered limits on distribution and density of maximum of mixed skew-t distribution under linear and power normalization, and further derived their higher-order expansions, respectively. Examples are given to support our findings.

math.ST↗
arXiv2016

Asymptotics and statistical inferences on independent and non-identically distributed bivariate Gaussian triangular arrays

In this paper, we establish the first and the second-order asymptotics of distributions of normalized maxima of independent and non-identically distributed bivariate Gaussian triangular arrays, where each vector of the $n$th row follows from a bivariate Gaussian distribution with correlation coefficient being a monotone continuous function of $i/n$. Furthermore, parametric inference for this unknown function is studied. Some simulation study and real data sets analysis are also presented.

stat.ME↗
arXiv2015

Second-order asymptotics for convolution of distributions with light tails

In this paper, asymptotic behavior of convolution of distributions belonging to two subclasses of distributions with exponential tails are considered, respectively. The precise second-order tail asymptotics of the convolutions are derived under the condition of second-order regular variation.

math.PR↗
arXiv2015

Dynamic Bivariate Normal Copula

Normal copula with a correlation coefficient between $-1$ and $1$ is tail independent and so it severely underestimates extreme probabilities. By letting the correlation coefficient in a normal copula depend on the sample size, Hüsler and Reiss (1989) showed that the tail can become asymptotically dependent. In this paper, we extend this result by deriving the limit of the normalized maximum of $n$ independent observations, where the $i$-th observation follows from a normal copula with its correlation coefficient being either a parametric or a nonparametric function of $i/n$. Furthermore, both parametric and nonparametric inference for this unknown function are studied, which can be employed to test the condition in Hüsler and Reiss (1989). A simulation study and real data analysis are presented too.

stat.ME↗
arXiv2012

Rates of convergence of extremes from skew normal samples

For a skew normal random sequence, convergence rates of the distribution of its partial maximum to the Gumbel extreme value distribution are derived. The asymptotic expansion of the distribution of the normalized maximum is given under an optimal choice of norming constants. We find that the optimal convergence rate of the normalized maximum to the Gumbel extreme value distribution is proportional to $1/\log n$.

stat.ME↗