A5029002117- Advanced Neural Network Applications
- Hand Gesture Recognition Systems
- Human Pose and Action Recognition
- Speech and Audio Processing
- Image Enhancement Techniques
- Advanced Vision and Imaging
- Image Processing Techniques and Applications
- Gait Recognition and Analysis
- Music and Audio Processing
- Industrial Vision Systems and Defect Detection
- Face recognition and analysis
- Water Quality Monitoring Technologies
- Remote-Sensing Image Classification
- Solar-Powered Water Purification Methods
- Fire Detection and Safety Systems
- Advanced Image Fusion Techniques
- Robot Manipulation and Learning
- Membrane Separation Technologies
- Anatomy and Medical Technology
- Recycling and Waste Management Techniques
- Advanced Chemical Sensor Technologies
- Medical Image Segmentation Techniques
- Surgical Simulation and Training
- Underwater Acoustics Research
- Constructed Wetlands for Wastewater Treatment
Xi'an Jiaotong University
2022-2024
Imperial College London
2023
Shandong University
2021
Southeast University
2020
Accurate instrument pose estimation is a crucial step towards the future of robotic surgery, enabling applications such as autonomous surgical task execution. Vision-based methods for provide practical approach to tool tracking, but they often require markers be attached instruments. Recently, more research has focused on development marker-less based deep learning. However, acquiring realistic data, with ground truth poses, required learning training, challenging. To address issues in...
In this study, we address the key challenges concerning accuracy and effectiveness of depth estimation for endoscopic imaging, with a particular emphasis on real-time inference impact light reflections. We propose novel lightweight solution named EndoDepthL that integrates Convolutional Neural Networks (CNN) Transformers to predict multi-scale maps. Our approach includes optimizing network architecture, incorporating dilated convolution, multi-channel attention mechanism. also introduce...
Illegal waste dumping not only encroaches on land resources but also threatens the health of surrounding residents. The traditional artificial monitoring solution requires professional workers to conduct field investigations. This high labor and economic costs demands a prolonged cycle for updating status. Therefore, some scholars use deep learning achieve automatic detection from satellite imagery. However, relevant models cannot effectively capture multi-scale features enhance key...
In this study, we address the key challenges concerning accuracy and effectiveness of depth estimation for endoscopic imaging, with a particular emphasis on real-time inference impact light reflections. We propose novel lightweight solution named EndoDepthL that integrates Convolutional Neural Networks (CNN) Transformers to predict multi-scale maps. Our approach includes optimizing network architecture, incorporating dilated convolution, multi-channel attention mechanism. also introduce...
Traditional speech enhancement uses only audio signals to obtain clean by removing the background noise. Audio-visual additional visual information improve intelligibility and perceptual quality of noisy speech, which can be applied video conferencing. However, original cues contain much redundant are very sensitive illumination. Meanwhile, related methods neglect mining crucial features fusion multimodal audio-visual features. In this paper, we propose an efficient feature network (MFF-Net)...