A5081233515- Robotics and Sensor-Based Localization
- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Advanced Vision and Imaging
- Domain Adaptation and Few-Shot Learning
- Image Enhancement Techniques
- Medical Image Segmentation Techniques
- Advanced Memory and Neural Computing
- Target Tracking and Data Fusion in Sensor Networks
- Speech and Audio Processing
- Nuclear reactor physics and engineering
- Underwater Acoustics Research
- Water Quality Monitoring Technologies
- Neuroscience and Neural Engineering
- Wireless Signal Modulation Classification
- Analog and Mixed-Signal Circuit Design
- Machine Learning in Materials Science
- 3D Surveying and Cultural Heritage
- Geophysical Methods and Applications
- Fusion materials and technologies
- EEG and Brain-Computer Interfaces
- Advanced X-ray and CT Imaging
- Advanced Image Processing Techniques
- Cerebrovascular and Carotid Artery Diseases
- Advanced SAR Imaging Techniques
Sun Yat-sen University
2021-2025
Southwest University of Science and Technology
2024-2025
Beijing Institute of Technology
2021-2024
Dalian University of Technology
2021
A myriad of recent breakthroughs in hand-crafted neural architectures for visual recognition have highlighted the urgent need to explore hybrid consisting diversified building blocks. Meanwhile, architecture search methods are surging with an expectation reduce human efforts. However, whether NAS can efficiently and effectively handle spaces disparate candidates (e.g. CNNs transformers) is still open question. In this work, we present Block-wisely Self-supervised Neural Architecture Search...
To boost the object grabbing capability of underwater robots for open-sea farming, we propose a new dataset (UDD) consisting three categories (seacucumber, seaurchin, and scallop) with 2,227 images. best our knowledge, it is first 4K HD collected in real farm. We also novel Poisson-blending Generative Adversarial Network (Poisson GAN) an efficient detection network (AquaNet) to address two common issues within related datasets: class-imbalance problem mass small object, respectively....
To achieve autonomous driving, developing 3D detection fusion methods, which aim to fuse the camera and LiDAR information, has draw great research interest in recent years. As a common practice, people rely on large-scale datasets fairly compare performance of different methods. While these have been carefully cleaned ideally minimize any potential noise, we observe that they cannot truly reflect data seen real vehicle, whose tends be noisy due various reasons. This hinders ability simply...
Recent advances in hand-crafted neural architectures for visual recognition underscore the pressing need to explore architecture designs comprising diverse building blocks. Concurrently, search (NAS) methods have gained traction as a means alleviate human efforts. Nevertheless, question of whether NAS can efficiently and effectively manage diversified spaces featuring disparate candidates, such Convolutional Neural Networks (CNNs) transformers, remains an open question. In this work, we...
While most recent autonomous driving system focuses on developing perception methods ego-vehicle sensors, people tend to overlook an alternative approach leverage intelligent roadside cameras extend the ability beyond visual range. We discover that state-of-the-art vision-centric detection perform poorly cameras. This is because these mainly focus recovering depth regarding camera center, where difference between car and ground quickly shrinks while distance increases. In this paper, we...
The micro-Doppler effect focuses on describing the detailed characteristics of moving targets and also plays a key role in field radar target recognition. In this paper, recurrent neural network (RNN) is used to classify signatures different targets. RNN models are sensitive temporal signals thus can learn necessary dependence signatures. This paper first constructs two-dimensional time-frequency distribution matrices by using short-time Fourier transformation (STFT). Then four types model...
3D multiple object tracking (MOT) plays a crucial role in autonomous driving perception. Recent end-to-end query-based trackers simultaneously detect and track objects, which have shown promising potential for the MOT task. However, existing methods overlook uncertainty issue, refers to lack of precise confidence about state location tracked objects. Uncertainty arises owing various factors during motion observation by cameras, especially occlusions small size target resulting an inaccurate...
A myriad of recent breakthroughs in hand-crafted neural architectures for visual recognition have highlighted the urgent need to explore hybrid consisting diversified building blocks. Meanwhile, architecture search methods are surging with an expectation reduce human efforts. However, whether NAS can efficiently and effectively handle spaces disparate candidates (e.g. CNNs transformers) is still open question. In this work, we present Block-wisely Self-supervised Neural Architecture Search...
In this paper, we propose a preliminary design for fusion-fission hybrid energy reactor (FFHER), based on current fusion science and technology well-developed fission technology. Design rules are listed primary concept blanket with uranium alloy as fuel water coolant is put forward. The can be natural uranium, LWR spent fuel, or depleted uranium. FFHER increase the utilization rate of in comparatively simple way to sustain development nuclear energy. interaction between neutron aim achieving...