A5079695819- Advanced Neural Network Applications
- Advanced SAR Imaging Techniques
- Remote Sensing and LiDAR Applications
- Robotics and Sensor-Based Localization
- Advanced Optical Sensing Technologies
- Radiomics and Machine Learning in Medical Imaging
- Video Surveillance and Tracking Methods
- Underwater Acoustics Research
- Maritime Navigation and Safety
- Underwater Vehicles and Communication Systems
- Domain Adaptation and Few-Shot Learning
- Lung Cancer Treatments and Mutations
- Advanced Image and Video Retrieval Techniques
- AI in cancer detection
- Digital Imaging for Blood Diseases
- Adversarial Robustness in Machine Learning
- Radar Systems and Signal Processing
- Automated Road and Building Extraction
- Multimodal Machine Learning Applications
- Autonomous Vehicle Technology and Safety
- Cancer Immunotherapy and Biomarkers
- Infrared Target Detection Methodologies
- Colorectal Cancer Treatments and Studies
- Advanced Semiconductor Detectors and Materials
- ECG Monitoring and Analysis
University of Liverpool
2022-2025
Jiangsu Industry Technology Research Institute
2023-2025
University of Hong Kong
2025
Hong Kong University of Science and Technology
2025
Xi’an Jiaotong-Liverpool University
2022-2024
University of Southampton
2021-2024
3D object detection is crucial for Autonomous Driving (AD) and Advanced Driver Assistance Systems (ADAS). However, most detectors prioritize accuracy, often overlooking network inference speed in practical applications. In this paper, we propose RadarNeXt, a real-time reliable detector based on the 4D mmWave radar point clouds. It leverages re-parameterizable neural networks to catch multi-scale features, reduce memory cost accelerate inference. Moreover, highlight irregular foreground...
The 3rd Workshop on Maritime Computer Vision (MaCVi) 2025 addresses maritime computer vision for Unmanned Surface Vehicles (USV) and underwater. This report offers a comprehensive overview of the findings from challenges. We provide both statistical qualitative analyses, evaluating trends over 700 submissions. All datasets, evaluation code, leaderboard are available to public at https://macvi.org/workshop/macvi25.
3D object detection and occupancy prediction are critical tasks in autonomous driving, attracting significant attention. Despite the potential of recent vision-based methods, they encounter challenges under adverse conditions. Thus, integrating cameras with next-generation 4D imaging radar to achieve unified multi-task perception is highly significant, though research this domain remains limited. In paper, we propose Doracamom, first framework that fuses multi-view for joint semantic...
4D millimeter-wave (MMW) radar, which provides both height information and dense point cloud data over 3D MMW has become increasingly popular in object detection. In recent years, radar-vision fusion models have demonstrated performance close to that of LiDAR-based models, offering advantages terms lower hardware costs better resilience extreme conditions. However, many treat radar as a sparse LiDAR, underutilizing radar-specific information. Additionally, these multi-modal networks are...
Current perception models for different tasks usually exist in modular forms on Unmanned Surface Vehicles (USVs), which infer extremely slowly parallel edge devices, causing the asynchrony between results and USV position, leading to error decisions of autonomous navigation. Compared with Ground (UGVs), robust USVs develops relatively slowly. Moreover, most current multi-task are huge parameters, slow inference not scalable. Oriented this, we propose Achelous, a low-cost fast unified...
Autonomous driving on water surfaces plays an essential role in executing hazardous and time-consuming missions, such as maritime surveillance, survivors rescue, environmental monitoring, hydrography mapping waste cleaning. This work presents WaterScenes, the first multi-task 4D radar-camera fusion dataset for autonomous surfaces. Equipped with a radar monocular camera, our Unmanned Surface Vehicle (USV) proffers all-weather solutions discerning object-related information, including color,...
Abstract Natural language (NL) based vehicle retrieval is a task aiming to retrieve that most consistent with given NL query from among all candidate vehicles. Because can be easily obtained, such has promising prospect in building an interactive intelligent traffic system (ITS). Current solutions mainly focus on extracting both text and image features mapping them the same latent space compare similarity. However, existing methods usually use dependency analysis or semantic role-labelling...
Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI
The aim of the present study was to predict response non-small cell lung cancer (NSCLC) patients immune checkpoint inhibitors (ICIs) by leveraging computed tomography (CT) images using deep learning techniques. Retrospectively, 624 sequential CT were gathered from 156 at Jiangsu Province Hospital, along with their clinical data. dataset subsequently partitioned into three groups: training (n=547), validation (n=64), and test (n=64). Moreover, an external cohort included 37 Nanjing Pukou...
With the rapid advancements of sensor technology and deep learning, autonomous driving systems are providing safe efficient access to intelligent vehicles as well transportation. Among these equipped sensors, radar plays a crucial role in robust perception information diverse environmental conditions. This review focuses on exploring different data representations utilized systems. Firstly, we introduce capabilities limitations by examining working principles signal processing measurements....
The perception of waterways based on human intent holds significant importance for autonomous navigation and operations Unmanned Surface Vehicles (USVs) in water environments. Inspired by visual grounding, this paper, we introduce WaterVG, the first grounding dataset designed USV-based waterway intention prompts. WaterVG encompasses prompts describing multiple targets, with annotations at instance level including bounding boxes masks. Notably, includes 11,568 samples 34,950 referred which...
In reality, images often exhibit multiple degradations, such as rain and fog at night (triple degradations). However, in many cases, individuals may not want to remove all for instance, a blurry lens revealing beautiful snowy landscape (double scenarios, people only desire deblur. These situations requirements shed light on new challenge image restoration, where model must perceive specific degradation types specified by human commands with degradations. We term this task Referring Flexible...