A5032223387- Adversarial Robustness in Machine Learning
- Visual Attention and Saliency Detection
- Bacillus and Francisella bacterial research
- Digital Media Forensic Detection
- Multimodal Machine Learning Applications
- Maritime Navigation and Safety
- Video Analysis and Summarization
- Underwater Vehicles and Communication Systems
- Olfactory and Sensory Function Studies
- Advanced Optical Sensing Technologies
- Advanced Image and Video Retrieval Techniques
- Advanced Neural Network Applications
- Anomaly Detection Techniques and Applications
- Robotic Path Planning Algorithms
- Ocular Surface and Contact Lens
- Integrated Circuits and Semiconductor Failure Analysis
Beijing Institute of Technology
2021-2024
Abstract The intrinsic similarity between camouflaged objects and background environment impedes the automatic detection/segmentation of objects, novel network architectures for deep learning are promising to overcome this challenge improve detection accuracy. However, these existing distinguishing their backgrounds do not account constraint speed, which results in high computational complexity inability meet requirements rapid detection. Therefore, based on human visual system, study...
Abstract The camouflaged object segmentation model (COSM) has recently gained substantial attention due to its remarkable ability detect objects. Nevertheless, deep vision models are widely acknowledged be susceptible adversarial examples, which can mislead models, causing them make incorrect predictions through imperceptible perturbations. vulnerability attacks raises significant concerns when deploying COSM in security-sensitive applications. Consequently, it is crucial determine whether...
It is necessary to plan a feasible path and avoid moving obstacles in real time, while the unmanned surface vehicle (USV) performs tasks archipelago environment. This paper focuses on obstacle avoidance of USV dynamic A hierarchical framework proposed for underactuated system USV. Devolution from planner controller more effective. Rapidly-exploring Random Trees Star used as global search path, combined with reactive approach modulate control input time. The feasibility method verified by...