A5017266884- Advanced Neural Network Applications
- Advanced Image and Video Retrieval Techniques
- Radiomics and Machine Learning in Medical Imaging
- Cancer-related Molecular Pathways
- CCD and CMOS Imaging Sensors
- Metal Alloys Wear and Properties
- Video Analysis and Summarization
- Optical Network Technologies
- Microtubule and mitosis dynamics
- Industrial Technology and Control Systems
- Cancer Immunotherapy and Biomarkers
- Metallurgy and Material Forming
- Cancer Mechanisms and Therapy
- Video Surveillance and Tracking Methods
- Immunotherapy and Immune Responses
- Simulation and Modeling Applications
- Cancer Genomics and Diagnostics
- Cell Image Analysis Techniques
- COVID-19 diagnosis using AI
- Iterative Learning Control Systems
- Advanced Memory and Neural Computing
- Error Correcting Code Techniques
- AI in cancer detection
- Advanced Wireless Communication Techniques
- Advanced Vision and Imaging
Guilin Medical University
2025
Ansteel (China)
2024
Wuhan University of Technology
2023
Heilongjiang University of Science and Technology
2022
Intel (United States)
2019
National Cheng Kung University
2019
There have been considerable debates over 2D and 3D representation learning on medical images. approaches could benefit from large-scale pretraining, whereas they are generally weak in capturing large contexts. natively strong contexts, however few publicly available dataset is diverse enough for universal pretraining. Even hybrid (2D + 3D) approaches, the intrinsic disadvantages within / parts still exist. In this study, we bridge gap between convolutions by reinventing convolutions. We...
The Corona Virus Disease 2019 (COVID-19) is highly infectious, has been spread worldwide, caused a global pandemic, and seriously endangered human health life. most effective methods for halting stopping the transmission of include early detection, quarantine, successful treatment. Because it exhibits significant imaging characteristics COVID-19 lesions in chest computed tomography (CT), can be used to diagnose COVID-19. Aiming at inaccuracies uneven gray distribution, irregular regions,...
The hot deformation behavior and microstructure evolution of a recently developed 2Cr12Ni4Mo3VNbN martensitic stainless steel are examined through compression tests conducted within the temperature range 900–1200 °C strain rate 0.01–10 s −1 . constitutive equation processing maps corresponding to established. activation energy for is determined be ≈457491.77 J mol Simultaneously, during studied. Based on analysis, it concluded that optimal windows 1106–1150 0.01–2.7 , as well at 1200 1–2.7...
Video instance segmentation (VIS) is a composite task that requires the joint detection, tracking, and of objects in video. In this work, we introduce complete framework for VIS, which integrates strengths general object tracking addressing unique challenges VIS. developing framework, investigate effective ways coordinating two components maximum benefits while thoroughly their separate contributions. Our approach improves over official baseline by an absolute 14.4% mAP achieves second place...
Cryo-em (Cryogenic electron microscopy) is a technology this can build bio-macromolecule of three-dimensional structure. Under the condition now, projection image biological macromolecule which collected by that contrast low, signal to noise blurring, and not easy distinguish single particle from background, corresponding processing lagging behind. Therefore, make Cryoem denoising useful, maintaining contour or function-construct improve quality resolution structure have important effect....
Immunotherapy shows durable response but only in a subset of patients, and test for predictive biomarkers requires procedures addition to routine workflow. We proposed confounder-aware representation learning-based system, genopathomic biomarker immunotherapy (PITER), that uses diagnosis-acquired hematoxylin-eosin (H&E)-stained pathological slides by leveraging histopathological genetic characteristics identify candidates immunotherapy. PITER was generated tested with three datasets...
In this paper, a new computer-aided diagnosis system is proposed to automatically diagnose liver cirrhosis based on fourphases CT images, which included non-contrast phase, arterial delay phase and portal venous phase. It developed for the purpose of discriminating into mild or severe level by automatic segmentation method classification using machine learning algorithm. First, gradient-inverse map images are calculated derive relative-smooth features in local area. Then we compared centroid...
Images captured by traditional cameras inevitably suffer from quality degradation under adverse conditions (e.g., low-light, snowy and foggy), this will jeopardize the safety of self-driving system. Unlike cameras, event can work stably in high dynamic range scenarios, capture accurate motion clues objects scenes. In paper, we propose a RGB-Event fusion framework to achieve robust vehicle detection degraded conditions. Firstly, design dual-stream feature extraction network extract...
On the basis of test and evaluation theory, software was developed to evaluate online dynamic performance CNC machine tool. After standard program run, response data obtained through collection system. Then tool fast appraised. Finally, all results were stored into database. It provided a good foundation for understanding real-time status variation rule
In the optical transport network, high-dimension equiprobable constellation is used to increase information transmission rate and forward error correction code improve system performance. However, adopting will lead some loss according Shannon theory performance of FEC needs be strengthened. Based on this, unequiprobable designed reduce by optimizing probability first bit taking value zero or one. order further performance, polar code, an advanced channel with good low decoding computation...