A5016700932- Multimodal Machine Learning Applications
- Human Pose and Action Recognition
- Advanced Image and Video Retrieval Techniques
- Domain Adaptation and Few-Shot Learning
- Anomaly Detection Techniques and Applications
- Retinal Imaging and Analysis
- Retinal Diseases and Treatments
- Image Retrieval and Classification Techniques
- Retinal and Macular Surgery
- Gait Recognition and Analysis
- Glaucoma and retinal disorders
- Diabetic Foot Ulcer Assessment and Management
- Text and Document Classification Technologies
Sun Yat-sen University
2021-2023
Suzhou University of Science and Technology
2022
University of Science and Technology of China
2021-2022
The task of image multi-label classification is to accurately recognize multiple objects in an input image. Most the recent works need leverage label co-occurrence matrix counted from training data construct graph structure, which are inflexible and may degrade model generalizability. In addition, these methods fail capture semantic correlation between channel feature maps further improve performance. To address issues, we propose DA-GAT (a D ouble A ttention framework based on G raph ne T...
Image multi-label classification task is mainly to correctly predict multiple object categories in the images. To capture correlation between labels, graph convolution network based methods have manually count label co-occurrence probability from training data construct a pre-defined as input of network, which inflexible and may degrade model generalizability. Moreover, most current cannot effectively align learned salient features with concepts, so that predicted results not be consistent...
Recent action localization works learn in a weakly supervised manner to avoid the expensive cost of human labeling. Those are mostly based on Multiple Instance Learning framework, where temporal pooling is an indispensable part that usually relies guidance snippet-level Class Activation Sequences (CAS) . However, we observe previous only leverage simple convolutional neural network for generation CAS, which ignores weak discriminative foreground segments and background ones, meanwhile,...
Retinal diseases and systemic diseases, such as diabetic retinopathy (DR) Alzheimer’s disease, may manifest themselves in the retina, changing retinal oxygen saturation ([Formula: see text]) level or vascular structures. Recent studies explored correlation of with either retina structures [Formula: text] level, but not both due to lack proper instrument methodology. In this study, we applied a dual-modal fundus camera developed deep learning-based analysis method simultaneously acquire...