A5068369601- Spine and Intervertebral Disc Pathology
- Medical Imaging and Analysis
- Infrared Target Detection Methodologies
- Video Surveillance and Tracking Methods
- Advanced Vision and Imaging
- Neurobiology and Insect Physiology Research
- CCD and CMOS Imaging Sensors
- Neural dynamics and brain function
- AI in cancer detection
- Aerospace Engineering and Energy Systems
- Speech and Audio Processing
- Advanced Memory and Neural Computing
- Visual Attention and Saliency Detection
- Bone Tissue Engineering Materials
- Domain Adaptation and Few-Shot Learning
- Image Enhancement Techniques
- Advanced Sensor and Energy Harvesting Materials
- Fault Detection and Control Systems
- Color Science and Applications
- Visual perception and processing mechanisms
- Advanced Image and Video Retrieval Techniques
- Speech Recognition and Synthesis
- Laser-Ablation Synthesis of Nanoparticles
- Neuroscience and Neural Engineering
- Advanced Sensor and Control Systems
Chinese Academy of Sciences
2024
Shenyang Institute of Automation
2024
Beijing University of Chemical Technology
2024
Ningxia University
2023
Guangdong University of Petrochemical Technology
2020-2023
University of Lincoln
2020-2022
Guangzhou University
2022
PLA Air Force Aviation University
2018
Lumbar vertebral body cancellous bone location and segmentation is crucial in an automated lumbar spine processing pipeline. Accurate reliable analysis of image expected to advantage practical medical diagnosis population-based strength. However, the design algorithms for demanding due significant anatomical variations scarcity publicly available data. In recent years, convolutional neural network (CNN) vision transformers (Vits) have been de facto standard segmentation. Although adept at...
Collision detection is critical for autonomous vehicles or robots to serve human society safely. Detecting looming objects robustly and timely plays an important role in collision avoidance systems. The locust lobula giant movement detector (LGMD1) specifically selective which are on a direct course. However, the existing LGMD1 models cannot distinguish object from near fast translatory moving object, because latter can evoke large amount of excitation that lead false spikes. This article...
Small moving objects at far distance always occupy only one or a few pixels in image and exhibit extremely limited visual features, which bring great challenges to motion detection. Highly evolved systems endow flying insects with remarkable ability pursue tiny mates prey, providing good template develop processing method for small target The insects' excellent sensitivity is believed come from class of specific neurons called detectors (STMDs). However, existing STMD-based methods often...
The LGMD1 neuron of locusts shows strong looming-sensitive property for both light and dark objects. Although a few models have been proposed, they are not reliable to inhibit the translating motion under certain conditions compare biological in locust. To address this issue, we propose bio-plausible model enhance collision-selectivity by inhibiting motion. proposed contains three parts, retina lamina layer receiving luminance change signals, medulla extracting cues via ON OFF pathways...
In the Weakly Supervised Object Detection (WSOD) task, detector is trained solely by image-level labels, which significantly reduced learning cost. However, in doing so, model lacks precise instance-level supervision, hinders its ability to accurately understand full object structure. As a result, tends focus on easily distinguishable local regions and make imprecise predictions. To address this problem, paper proposed Part Whole (P2W) approach, consisting of two main components:...
Human motion capture is widely applied in engineering bionics, ergonomics, aerospace, military defense, biomedical, sports competition and other fields.However, the human equipments are less developed China.This paper presents a system based on computer vision using binocular camera.Stereo camera calibration an important step vision.It occupies status stereo distance measurement, three-dimensional reconstruction fields.In this paper, new method of proposed monocular camera.It can also get...
Abstract It is an enormous challenge for intelligent vehicles to avoid collision accidents at night because of the extremely poor light conditions. Thermal cameras can capture temperature map night, even with no sources and are ideal detection in darkness. However, how extract cues efficiently effectively from captured limited computing resources still a key issue be solved. Recently, bio-inspired neural network LGMD has been proposed successfully, but daytime visible light. Whether it used...
In low light conditions, image enhancement is critical for vision-based artificial systems since details of objects in dark regions are buried. Moreover, enhancing the low-light without introducing too many irrelevant artifacts important visual tasks like motion detection. However, conventional methods always have risk "bad" enhancement. Nocturnal insects show remarkable abilities at night time, and their adaptations responses provide inspiration this paper, we aim to adopt neural mechanism...