papers

Publications (27)

cs.CV2025

AD^2-Bench: A Hierarchical CoT Benchmark for MLLM in Autonomous Driving under Adverse Conditions

Zhaoyang Wei, Chenhui Qiang, Bowen Jiang +3

Chain-of-Thought (CoT) reasoning has emerged as a powerful approach to enhance the structured, multi-step decision-making capabilities of Multi-Modal Large Models (MLLMs), is parti…

cs.CV2025

Boosting Segment Anything Model Towards Open-Vocabulary Learning

Xumeng Han, Longhui Wei, Xuehui Yu +6

The recent Segment Anything Model (SAM) has emerged as a new paradigmatic vision foundation model, showcasing potent zero-shot generalization and flexible prompting. Despite SAM fi…

cs.CV2022

Object Localization under Single Coarse Point Supervision

Xuehui Yu, Pengfei Chen, Di Wu +6

Point-based object localization (POL), which pursues high-performance object sensing under low-cost data annotation, has attracted increased attention. However, the point annotatio…

cs.AI2026

Conflict-Aware Additive Guidance for Flow Models under Compositional Rewards

Xuehui Yu, Fucheng Cai, Meiyi Wang +2

Inference-time guided sampling steers state-of-the-art diffusion and flow models without fine-tuning by interpreting the generation process as a controllable trajectory. This provi…

cs.CV2024

Rethinking Sampling Strategies for Unsupervised Person Re-identification

Xumeng Han, Xuehui Yu, Guorong Li +5

Unsupervised person re-identification (re-ID) remains a challenging task. While extensive research has focused on the framework design and loss function, this paper shows that samp…

cs.CV2022

P2P-Loc: Point to Point Tiny Person Localization

Xuehui Yu, Di Wu, Qixiang Ye +2

Bounding-box annotation form has been the most frequently used method for visual object localization tasks. However, bounding-box annotation relies on a large amount of precisely a…

cs.CV2024

ClickTrack: Towards Real-time Interactive Single Object Tracking

Kuiran Wang, Xuehui Yu, Wenwen Yu +5

Single object tracking(SOT) relies on precise object bounding box initialization. In this paper, we reconsidered the deficiencies in the current approaches to initializing single o…

cs.LG2024

Causal prompting model-based offline reinforcement learning

Xuehui Yu, Yi Guan, Rujia Shen +3

Model-based offline Reinforcement Learning (RL) allows agents to fully utilise pre-collected datasets without requiring additional or unethical explorations. However, applying mode…

cs.CV2026

Towards Any-Quality Image Segmentation via Generative and Adaptive Latent Space Enhancement

Guangqian Guo, Aixi Ren, Yong Guo +6

Segment Anything Models (SAMs), known for their exceptional zero-shot segmentation performance, have garnered significant attention in the research community. Nevertheless, their p…

cs.CV2025

ReLayout: Integrating Relation Reasoning for Content-aware Layout Generation with Multi-modal Large Language Models

Jiaxu Tian, Xuehui Yu, Yaoxing Wang +3

Content-aware layout aims to arrange design elements appropriately on a given canvas to convey information effectively. Recently, the trend for this task has been to leverage large…

cs.CV2026

Segment Any-Quality Images with Generative Latent Space Enhancement

Guangqian Guo, Yong Guo, Xuehui Yu +3

Despite their success, Segment Anything Models (SAMs) experience significant performance drops on severely degraded, low-quality images, limiting their effectiveness in real-world…

cs.CV2023

Spatial Self-Distillation for Object Detection with Inaccurate Bounding Boxes

Di Wu, Pengfei Chen, Xuehui Yu +3

Object detection via inaccurate bounding boxes supervision has boosted a broad interest due to the expensive high-quality annotation data or the occasional inevitability of low ann…

cs.CV2026

SAPNet++: Evolving Point-Prompted Instance Segmentation with Semantic and Spatial Awareness

Zhaoyang Wei, Xumeng Han, Xuehui Yu +4

Single-point annotation is increasingly prominent in visual tasks for labeling cost reduction. However, it challenges tasks requiring high precision, such as the point-prompted ins…

cs.CV2021

SM+: Refined Scale Match for Tiny Person Detection

Nan Jiang, Xuehui Yu, Xiaoke Peng +2

Detecting tiny objects ( e.g., less than 20 x 20 pixels) in large-scale images is an important yet open problem. Modern CNN-based detectors are challenged by the scale mismatch bet…

cs.CV2025

P2Object: Single Point Supervised Object Detection and Instance Segmentation

Pengfei Chen, Xuehui Yu, Xumeng Han +5

Object recognition using single-point supervision has attracted increasing attention recently. However, the performance gap compared with fully-supervised algorithms remains large.…

cs.AI2022

Causal Coupled Mechanisms: A Control Method with Cooperation and Competition for Complex System

Xuehui Yu, Jingchi Jiang, Xinmiao Yu +2

Complex systems are ubiquitous in the real world and tend to have complicated and poorly understood dynamics. For their control issues, the challenge is to guarantee accuracy, robu…

cs.CV2024

P2RBox: Point Prompt Oriented Object Detection with SAM

Guangming Cao, Xuehui Yu, Wenwen Yu +5

Single-point annotation in oriented object detection of remote sensing scenarios is gaining increasing attention due to its cost-effectiveness. However, due to the granularity ambi…

cs.CV2024

CPR++: Object Localization via Single Coarse Point Supervision

Xuehui Yu, Pengfei Chen, Kuiran Wang +5

Point-based object localization (POL), which pursues high-performance object sensing under low-cost data annotation, has attracted increased attention. However, the point annotatio…

cs.CV2020

Effective Fusion Factor in FPN for Tiny Object Detection

Yuqi Gong, Xuehui Yu, Yao Ding +3

FPN-based detectors have made significant progress in general object detection, e.g., MS COCO and PASCAL VOC. However, these detectors fail in certain application scenarios, e.g.,…

cs.CV2024

P2Seg: Pointly-supervised Segmentation via Mutual Distillation

Zipeng Wang, Xuehui Yu, Xumeng Han +4

Point-level Supervised Instance Segmentation (PSIS) aims to enhance the applicability and scalability of instance segmentation by utilizing low-cost yet instance-informative annota…

cs.CV2019

Scale Match for Tiny Person Detection

Xuehui Yu, Yuqi Gong, Nan Jiang +2

Visual object detection has achieved unprecedented ad-vance with the rise of deep convolutional neural networks.However, detecting tiny objects (for example tiny per-sons less than…

cs.CV2022

Point-to-Box Network for Accurate Object Detection via Single Point Supervision

Pengfei Chen, Xuehui Yu, Xumeng Han +7

Object detection using single point supervision has received increasing attention over the years. However, the performance gap between point supervised object detection (PSOD) and…

cs.LG2024

Skill-aware Mutual Information Optimisation for Generalisation in Reinforcement Learning

Xuehui Yu, Mhairi Dunion, Xin Li +1

Meta-Reinforcement Learning (Meta-RL) agents can struggle to operate across tasks with varying environmental features that require different optimal skills (i.e., different modes o…

cs.CV2024

Semantic-aware SAM for Point-Prompted Instance Segmentation

Zhaoyang Wei, Pengfei Chen, Xuehui Yu +3

Single-point annotation in visual tasks, with the goal of minimizing labelling costs, is becoming increasingly prominent in research. Recently, visual foundation models, such as Se…

cs.CV2020

The 1st Tiny Object Detection Challenge:Methods and Results

Xuehui Yu, Zhenjun Han, Yuqi Gong +22

The 1st Tiny Object Detection (TOD) Challenge aims to encourage research in developing novel and accurate methods for tiny object detection in images which have wide views, with a…

cs.CV2021

Anti-UAV: A Large Multi-Modal Benchmark for UAV Tracking

Nan Jiang, Kuiran Wang, Xiaoke Peng +7

Unmanned Aerial Vehicle (UAV) offers lots of applications in both commerce and recreation. With this, monitoring the operation status of UAVs is crucially important. In this work,…

cs.CV2025

SAM-CP: Marrying SAM with Composable Prompts for Versatile Segmentation

Pengfei Chen, Lingxi Xie, Xinyue Huo +5

The Segment Anything model (SAM) has shown a generalized ability to group image pixels into patches, but applying it to semantic-aware segmentation still faces major challenges. Th…