Xuezhi Xiang

tcnerv:dual-domain temporal context modeling for implicit neural video compression

Xuezhi Xiang, Yixin Zhao, Heqi Xiang, Jiayao Liu et autres

Video compression aims to minimize reconstruction distor tion under a constrained bit rate. Existing video implicit neural representations (INRs) often decode frames independently, leaving intermediate features unconditioned on previous reconstructions and content embeddings without explicit temporal prediction. We propose TCNeRV, which exploits …

PSMP-CLIP: Patch-Prompt SAM and Multi-Semantic Prompting for CLIP-Based Zero-Shot Anomaly Detection

Xuezhi Xiang, Guanghao Wu, Heqi Xiang, Jiayao Liu et autres

Zero-shot anomaly detection aims to localize anomalies without target-domain samples. Existing CLIP-based methods suffer from coarse anomaly maps and limited semantic prompts. We propose PSMP-CLIP, integrating patch-prompt SAM2 segmentation (PPSS) and multi-semantic guided prompt regularization (MSGPR). PPSS samples prompts directly from intermediate …

PVFT-Net: A point-voxel fusion method for self-supervised scene flow estimation with transformer

Xuezhi Xiang, Xi Wang, Xiaoheng Li, Xiankun Zhou et autres

cn (code pays fourni par la source)

Mamba-SF: Monocular Scene Flow Learning with State Space Models

Yiming Chen, Xuezhi Xiang, Xianye Ben, Insha Hassan et autres

Monocular scene flow estimation has been a long-standing problem in computer vision. Methods based on the RAFT architecture are currently the mainstream approaches, while often overlooking the long-range dependencies in motion and texture features and fail to fully utilize the spatial information …

cn (code pays fourni par la source)

Scene flow estimation with local surface representation

Xi Wang, Xuezhi Xiang, Denis Ombati Omweri, Himaloy Himu

Scene flow estimation aims to compute the 3D displacement field across consecutive point cloud frames and finds extensive applications in the 3D perception field. Most current methods extract features only through xyz coordinates, which lacks sufficient local context information, making it difficult …

cn (code pays fourni par la source)

CDRFormer: CNN and dimensional reciprocal attention mixing transformer network for remote sensing images semantic segmentation

Xuezhi Xiang, Denis Ombati Omweri, Himaloy Himu

Semantic segmentation of remote sensing images is a fundamental task in geospatial research. Due to the complex ground objects, rich feature details, large intraclass variance, and small interclass variance in remote sensing images, deep learning semantic segmentation methods are usually required to …

cn (code pays fourni par la source)

Scene Flow Estimation for Autonomous Driving via Correlation Compensation and Initial Motion Check

Mingliang Zhai, Bing‐Kun Bao, Xuezhi Xiang

Scene flow estimation from LiDAR sensors is a crucial task for dynamic environmental perception in autonomous driving scenarios. Recently, self-supervised approaches have gained attention for their ability to reduce the burden of point-wise annotation. Although existing methods have been able to generate …

cn (code pays fourni par la source)

Deep scene flow learning from point cloud with Transformer

Xuezhi Xiang, Yiming Chen, Rokia Abdein, Lei Zhang et autres

cn (code pays fourni par la source)

LKA-ReID:Vehicle Re-Identification with Large Kernel Attention

Xuezhi Xiang, Zongmin Ma, Lei Zhang, Denis Ombati et autres

With the rapid development of intelligent transportation systems and the popularity of smart city infrastructure, Vehicle Re-ID technology has become an important research field. The vehicle Re-ID task faces an important challenge, which is the high similarity between different vehicles. Existing methods …

cn (code pays fourni par la source)

Self-supervised monocular depth estimation with large kernel attention and dynamic scene perception

Xuezhi Xiang, Yao Wang, Xiaoheng Li, Lei Zhang et autres

cn (code pays fourni par la source)

DMRFlow: 4D Radar Scene Flow Estimation With Decoupled Matching and Refinement

Mingliang Zhai, Bing-Kun Bao, Xuezhi Xiang

Scene flow estimation from 4D radar sensors has become increasingly popular in recent years. In this paper, we propose a matching and refinement decoupling method to estimate scene flow from 4D radar point clouds. Since 4D radar point clouds are much sparser …

cn (code pays fourni par la source)

Vehicle re-identification with large separable kernel attention and hybrid channel attention

Xuezhi Xiang, Zongmin Ma, Xiaoheng Li, Lei Zhang et autres

cn (code pays fourni par la source)