A5038786829- Advanced Image Fusion Techniques
- Smart Agriculture and AI
- Advanced Image Processing Techniques
- Spectroscopy and Chemometric Analyses
- Orthopedic Surgery and Rehabilitation
- Astrophysics and Star Formation Studies
- Image and Video Quality Assessment
- Stellar, planetary, and galactic studies
- Remote Sensing and LiDAR Applications
- Nematode management and characterization studies
- Diabetic Foot Ulcer Assessment and Management
- Shoulder Injury and Treatment
- 3D Surveying and Cultural Heritage
- Astronomy and Astrophysical Research
- Image and Signal Denoising Methods
- Advanced Vision and Imaging
- Foot and Ankle Surgery
- Gamma-ray bursts and supernovae
- Date Palm Research Studies
- Advanced Chemical Sensor Technologies
- Winter Sports Injuries and Performance
Ningbo Institute of Industrial Technology
2021-2023
Chinese Academy of Sciences
2021-2023
University of Chinese Academy of Sciences
2022-2023
University of California, Berkeley
2023
We measure the 3D kinematic structures of young stars within central 0.5 parsec our Galactic Center using 10 m telescopes W.~M.~Keck Observatory over a time span 25 years. Using high-precision measurements positions on sky, and proper motions radial velocities from new observations literature, we constrain orbital parameters for each star. Our results show two statistically significant sub-structures: clockwise stellar disk with 18 candidate stars, as has been proposed before, but an...
With the development of image recovery models, especially those based on adversarial and perceptual losses, detailed texture portions images are being recovered more naturally. However, these restored similar but not identical in detail to their reference images. traditional quality assessment methods, results with better subjective perceived often score lower objective scoring. Assessment methods suffer from inconsistencies. This paper proposes a regional differential information entropy...
This paper proposed a method for reconstructing floorplans from indoor point clouds. Unlike existing corner and line primitive detection algorithms, this uses generative adversarial network to learn the complex distribution of layout graphics, repairs incomplete room masks into more regular segmentation areas. Automatic learning structure information graphics can reduce dependence on geometric priors, replacing optimization algorithms with Deep Neural Networks (DNN) improve efficiency data...
High-precision, high-speed detection and classification of weakly differentiated targets has always been a difficult problem in the field image vision. In this paper, phytopathogenic Bursaphelenchus xylophilus with small size very weak inter-species differences is taken as an example. Our work aimed at current target detection: We propose lightweight self attention network. Experiments show that key feature recognition areas plant nematodes found by our Self Attention network are good...
We measure the 3D kinematic structures of young stars within central 0.5 parsec our Galactic Center using 10 m telescopes W.~M.~Keck Observatory over a time span 25 years. Using high-precision measurements positions on sky, and proper motions radial velocities from new observations literature, we constrain orbital parameters for each star. Our results show two statistically significant sub-structures: clockwise stellar disk with 18 candidate stars, as has been proposed before, but an...
Angular measurements are essential for an appropriate treatment Hallux valgus (HV), a common forefoot deformity. However, it still depends on manual labeling and measurement, which is time-consuming sometimes unreliable. Automating this process thing of concern. lacks dataset, the key points-based method that succeeded in pose estimation not suitable angle estimation. In paper, We make dataset develop algorithm based image segmentation linear regression to solve problems. It shows great...
Image denoising is a fundamental problem in computer vision and has received much attention from scholars. With the fast development of convolutional neural networks, more deep learning-based noise reduction algorithms have emerged. However, current image networks tend to apply only RGB color space, ignoring information at visual perception level, making images generated by these too smooth lacking texture details. Therefore, this paper proposes novel network translation area using learning...
Bursaphelenchus is the key quarantine object of customs department, but there are many kinds and their morphological changes very complicated. Therefore, identification has long been a difficult problem in work various customs. We propose new method based on convolutional neural network with pluggable attention module. The first uses to extract features Bursaphelenchus, then obtains map characteristic part through module, adds for training, strengthens different region weights, thereby...
Abstract Background: High-precision, high-speed detection and classification of weakly differentiated targets has always been a difficult problem in the field image vision. In this paper, phytopathogenic Bursaphelenchus xylophilus with small size very weak inter-species differences is taken as an example. Methods: Our work carried out response to current target problems: a. To replace complex network labelling training process based on expert empirical knowledge, we proposed lightweight...
With the development of image recovery models,especially those based on adversarial and perceptual losses,the detailed texture portions images are being recovered more naturally.However,these restored similar but not identical in detail to their reference images.With traditional quality assessment methods,results with better subjective perceived often score lower objective scoring.Assessment methods suffer from inconsistencies.This paper proposes a regional differential information entropy...
Angular measurements is essential to make a resonable treatment for Hallux valgus (HV), common forefoot deformity. However, it still depends on manual labeling and measurement, which time-consuming sometimes unreliable. Automating this process thing of concern. lack dataset the keypoints based method made great success in pose estimation not suitable field.To solve problems, we developed an algorithm deep learning linear regression. It shows fitting ability ground truth.