Recent research on super-resolution has achieved great success due to the development of deep convolutional neural networks (DCNNs). However, arbitrary scale factor been ignored for a long time. Most previous researchers regard differentscale factors as independent tasks. They train specific model each which is inefficient in computing, and prior work only take several integer into consideration. In this work,we propose novel method called Meta-SR firstly solve (including non-integer...

Gait recognition is beneficial for a variety of applications, including video surveillance, crime scene investigation, and social security, to mention few. However, gait often suffers from multiple exterior factors in real scenes, such as carrying conditions, wearing overcoats, diverse viewing angles. Recently, various deep learning-based methods have achieved promising results, but they tend extract one the salient features using fixed-weighted convolutional networks, do not well consider...

Recent works on pose-based gait recognition have demonstrated the potential of using such simple information to achieve results comparable silhouette-based methods. However, generalization ability methods different datasets is undesirably inferior that ones, which has received little attention but hinders application these in real-world scenarios. To improve across datasets, we propose a Generalized Pose-based Gait (GPGait) framework. First, Human-Oriented Transformation (HOT) and series...

The performance of salient object segmentation has been significantly advanced by using the deep convolutional networks. However, these networks often produce blob-like saliency maps without accurate boundaries. This is caused limited spatial resolution their feature after multiple pooling operations and might hinder downstream applications that require precise shapes. To address this issue, we propose a novel model—Focal Boundary Guided (Focal-BG) network. Our model designed to jointly...

Shadow detection is an important and challenging problem in computer vision. Recently, single image shadow had achieved major progress with the development of deep convolutional networks. However, existing methods are still vulnerable to background clutters, often fail capture global context input image. These contextual semantic cues essential for accurately localizing regions. Moreover, rich spatial details required segment regions precise shape. To this end, paper presents a novel model...

Recent research on super-resolution has achieved great success due to the development of deep convolutional neural networks (DCNNs). However, arbitrary scale factor been ignored for a long time. Most previous researchers regard different factors as independent tasks. They train specific model each which is inefficient in computing, and prior work only take several integer into consideration. In this work, we propose novel method called Meta-SR firstly solve (including non-integer factors)...

Recently, diffusion-based methods for monocular 3D human pose estimation have achieved state-of-the-art (SOTA) performance by directly regressing the joint coordinates from 2D sequence. Although some decompose task into bone length and direction prediction based on anatomical skeleton to explicitly incorporate more body prior constraints, of these is significantly lower than that SOTA methods. This can be attributed tree structure skeleton. Direct application disentangled method could...

Recent research on single image super-resolution (SISR) has achieved great success due to the development of deep convolutional neural networks. However, most existing SISR methods merely focus a fixed integer scale factor. This simplified assumption does not meet complex conditions for real-world images which often suffer from various blur kernels or levels noise. More importantly, previous lack ability cope with arbitrary degradation parameters (scale factors, kernels, and noise levels)...

Image saliency detection has recently witnessed rapid progress due to deep convolutional neural networks. However, the typical binary cross entropy loss used in networks by is a pixel-wise loss, resulting independent prediction of salient probability each pixel. It raises problem spatial discontinuity predicted maps. Many researchers try solve this using super-pixel segmentation, but it complicated and time-consuming. In paper, we propose an Adversarial Saliency Detection Network (ASDN)...

Recently, diffusion-based methods for monocular 3D human pose estimation have achieved state-of-the-art (SOTA) performance by directly regressing the joint coordinates from 2D sequence. Although some decompose task into bone length and direction prediction based on anatomical skeleton to explicitly incorporate more body prior constraints, of these is significantly lower than that SOTA methods. This can be attributed tree structure skeleton. Direct application disentangled method could...

Gait recognition plays an important role in video surveillance and security by identifying humans based on their unique walking patterns. The existing gait methods have achieved competitive accuracy with shape motion patterns under limited-covariate conditions. However, when extreme appearance changes distort discriminative features, yields unsatisfactory results cross-covariate In this work, we first indicate that the integral pose each silhouette maintains appearance-unrelated identity....

Generating dances that are both lifelike and well-aligned with music continues to be a challenging task in the cross-modal domain. This paper introduces PopDanceSet, first dataset tailored preferences of young audiences, enabling generation aesthetically oriented dances. And it surpasses AIST++ genre diversity intricacy depth dance movements. Moreover, proposed POPDG model within iDDPM framework enhances and, through Space Augmentation Algorithm, strengthens spatial physical connections...

Recent works on pose-based gait recognition have demonstrated the potential of using such simple information to achieve results comparable silhouette-based methods. However, generalization ability methods different datasets is undesirably inferior that ones, which has received little attention but hinders application these in real-world scenarios. To improve across datasets, we propose a \textbf{G}eneralized \textbf{P}ose-based \textbf{Gait} (\textbf{GPGait}) framework. First, Human-Oriented...

Depression, a highly prevalent mental illness, affects over 280 million individuals worldwide. Early detection and timely intervention are crucial for promoting remission, preventing relapse, alleviating the emotional financial burdens associated with depression. However, patients depression often go undiagnosed in primary care setting. Unlike many physiological illnesses, lacks objective indicators recognizing risk, existing methods risk recognition time-consuming encounter shortage of...