A5068335881- Advanced Radiotherapy Techniques
- Advanced Neural Network Applications
- Radiomics and Machine Learning in Medical Imaging
- Medical Imaging Techniques and Applications
- Machine Learning and ELM
- Digital Radiography and Breast Imaging
- Domain Adaptation and Few-Shot Learning
- Advanced Memory and Neural Computing
- Global Cancer Incidence and Screening
- Radiation Therapy and Dosimetry
- Head and Neck Cancer Studies
- COVID-19 diagnosis using AI
- CCD and CMOS Imaging Sensors
- Image and Video Stabilization
- Advanced Vision and Imaging
- Advanced Image Processing Techniques
- Advanced Image and Video Retrieval Techniques
- Multimodal Machine Learning Applications
- Advanced X-ray and CT Imaging
- AI in cancer detection
Guilin University of Technology
2023-2024
Fourth Affiliated Hospital of Guangxi Medical University
2018-2022
We developed a deep learning model to achieve automatic multitarget delineation on planning CT (pCT) and synthetic (sCT) images generated from cone-beam (CBCT) images. The geometric dosimetric impact of the was evaluated for breast cancer adaptive radiation therapy.We retrospectively analyzed 1,127 patients treated with radiotherapy after breast-conserving surgery two medical institutions. CBCT patient setup acquired utilizing breath-hold guided by optical surface monitoring system were used...
Accurate segmentation of gross target volume (GTV) from computed tomography (CT) images is a prerequisite in radiotherapy for nasopharyngeal carcinoma (NPC). However, this task very challenging due to the low contrast at boundary tumor and great variety sizes morphologies tumors between different stages. Meanwhile, data source also seriously affect results segmentation. In paper, we propose novel three-dimensional (3D) automatic algorithm that adopts cascaded multiscale local enhancement...
In recent years, edge intelligence (EI) has emerged, combining computing with AI, and specifically deep learning, to run AI algorithms directly on devices. practical applications, EI faces challenges related computational power, power consumption, size, cost, the primary challenge being trade-off between consumption. This rendered traditional platforms unsustainable, making heterogeneous parallel a crucial pathway for implementing EI. our research, we leveraged Xilinx Zynq 7000 platform,...
In recent years, Edge Intelligence (EI) has emerged, combining edge computing with AI, specifically deep learning, to run AI algorithms directly on devices. practical applications, EI faces challenges related computational power, power consumption, size, and cost, the primary challenge being trade-off between consumption. This rendered traditional platforms unsustainable, making heterogeneous parallel a crucial pathway for implementing EI. our research, we leveraged Xilinx Zynq 7000...
In computer vision, image recognition is one of the classic tasks. Currently, with foundation big data and advanced hardware, deep learning has achieved high accuracy. However, often fails to perform well when faced a small number samples. Therefore, few-shot become key technology solve this problem. The paradigm different from that learning. It aims learn universal representation multiple training categories, used for in new categories. Each instance consists group images an unlabeled...
Objective To analyze the relationship between planning factors of intensity-modulated radiation therapy (IMRT) and gamma index investigate effect each parameter upon γ passing rate IMRT. Methods Gamma analysis was performed using 3%/3 mm acceptance criteria for 457 IMRT beams with different factors. During multi-factor ANOVA rate, control variables primarily included minimum segment area, number monitor unit (MU), segment, conversation, spatial resolution in measured dose...
The main challenge of few-shot learning lies in the limited labeled sample data. In addition, since image-level labels are usually not accurate describing features images, it leads to difficulty for model have good generalization ability and robustness. This problem has been well solved yet, existing metric-based methods still room improvement. To address this issue, we propose a method based on three-dimension attention mechanism self-supervised learning. module is used extract more...
Objective To compare the dosimetric characteristics of VMAT plans between flattening-filter-free (FFF) and flattening filter (FF) modes for nasopharyngeal carcinoma analyze feasibility in FFF model applied clinical practice. Methods Ten patients diagnosed with stage Ⅱ 2016 2017 were recruited this investigation. For CT image target volume (CTV) identical patient, FFF-and FF-mode established prescription dose 6 975 cGy modified parameters on RayStation platform (6 MV X-ray). The...
Objective To evaluate the dosimetric effects of set-up errors on nasal NK/T cell lymphoma by introducing into radiotherapy planning system for dose reconstruction. Methods Ten patients with were recruited. A non-coplanar volumetric modulated arc therapy plan was designed CT image and clinical target area each patient. After completion plan, introduced changing ISO treatment, calculation performed to reconstruct distribution. Results With increase errors, decreased order affected...