A5019114101- Natural Language Processing Techniques
- Video Surveillance and Tracking Methods
- Topic Modeling
- Face recognition and analysis
- Medical Image Segmentation Techniques
- Radiomics and Machine Learning in Medical Imaging
- Digital Media and Visual Art
- Hydrology and Watershed Management Studies
- Advanced Neural Network Applications
- Sleep and Work-Related Fatigue
- Gaze Tracking and Assistive Technology
- Flood Risk Assessment and Management
- Advanced Image and Video Retrieval Techniques
- COVID-19 diagnosis using AI
- Data Quality and Management
- Image Retrieval and Classification Techniques
- Advanced MRI Techniques and Applications
- Hydrological Forecasting Using AI
- Pancreatic and Hepatic Oncology Research
- AI in cancer detection
- Biomedical Text Mining and Ontologies
- Brain Tumor Detection and Classification
- Gastric Cancer Management and Outcomes
- Colorectal Cancer Screening and Detection
- Human Pose and Action Recognition
China Aerospace Science and Industry Corporation (China)
2021-2022
China Aerospace Science and Technology Corporation
2022
Shanghai Institute of Technology
2022
Waseda University
2022
Beijing University of Posts and Telecommunications
2019-2021
University of International Relations
2020
North China University of Water Resources and Electric Power
2020
University of Pittsburgh
2000
The word embedding of traditional named entity recognition (NER) methods can't represent the polysemy a word, fully consider contextual information, and local features in extraction process are easy to be ignored. To solve this problem, method based on Bert is proposed, BERT-BiLSTM-IDCNN-CRF model built. used for pre-training, then trained vectors input into bidirectional long-term short-term memory network (BiLSTM) iterative dilation convolution (IDCNN) extract features. Then output two...
Gastrointestinal (GI) endoscopy has been an active field of research motivated by the large number highly lethal GI cancers. Early cancer precursors are often missed during endoscopic surveillance. The high rate such abnormalities is thus a critical bottleneck. Lack attentiveness due to tiring procedures, and requirement training few contributing factors. An automatic disease classification system can help reduce risks flagging suspicious frames lesions. consists several multi-organ...
Abstract Due to the nonlinear characteristics of runoff data and poor performance single prediction model, a weighted integrated modified complementary ensemble empirical mode decomposition (MCEEMD)-based model was proposed predict monthly three hydrological stations in lower reaches Yellow River. In this particle swarm optimization (PSO) used optimize parameters support vector regression (SVR), back propagation neural network (BP), long short-term memory (LSTM) that constitute it. The...
Medical image classification and diagnosis is currently a hot topic in the field of deep learning. The ACM International Conference on Multimedia Simula co-hosted MutilMedia Grand Challenge, which aims to use artificial intelligence aiding detection gastrointestinal image. This competition divided into four subtasks, including detection, efficient (the same hardware for all participants) report generation. We participate multi-label task propose new attention model, can effectively improve...
Video identification is an important task in the practical application and industry. Based on iQIYI-VID-2019 dataset, ACM International Conference Multimedia iQIYI co-hosted celebrity video challenge. We take part competition, propose a new feature fusion method design residual dense network which can improve performance complex scenes. Only with face features, we achieve 0.9035 mean Average Precision(mAP) win second place leadboard. At same time, it best score only official features. It...
In this busy modern society, there are many external and psychological factors that can cause people to feel tired. The severity of fatigue driving is comparable drunk when we consider the accident rate. Therefore, how avoid situation has become an important issue. With trend machine learning becoming more mature, facial expression recognition been widely used in real life. A large number studies reports about detection improve be found. Most them either detect drowsiness states without...
Entity recognition and relation extraction techniques in text mining are very important for the of sensitive information contract texts. Due to lack corpus field texts, existing entity models cannot extract relationships based on semantics In view above problems, order enable model apply domain characteristics knowledge domain, this paper proposes a joint entity-relation method that integrates representation features, incorporates dictionary features into encoder part through self-attention...
Aiming at the problem that sampling parts need to be manually marked in process of facial skin quality detection, a detection system based on key points face is proposed automatically locate parts. In order realize real-time and fast portable devices, adopts an improved lightweight MobileNetV2 network. Adjusted model structure make parameters smaller improve rate. At same time, accuracy network, efficient channel attention (ECA) module added basis inverted residual block, which solves...