A5102446062- Advanced Radiotherapy Techniques
- Radiomics and Machine Learning in Medical Imaging
- Radiation Therapy and Dosimetry
- Algebraic Geometry and Number Theory
- Homotopy and Cohomology in Algebraic Topology
- AI in cancer detection
- Advances in Oncology and Radiotherapy
- Management of metastatic bone disease
- Advanced X-ray and CT Imaging
- Brain Tumor Detection and Classification
- Spine and Intervertebral Disc Pathology
- Advanced Neural Network Applications
- MRI in cancer diagnosis
- Topic Modeling
- Lung Cancer Diagnosis and Treatment
- Brain Metastases and Treatment
- Medical Imaging Techniques and Applications
- Geometry and complex manifolds
- Colorectal and Anal Carcinomas
- Cloud Computing and Resource Management
- Sarcoma Diagnosis and Treatment
- Natural Language Processing Techniques
- Adversarial Robustness in Machine Learning
- Digital Imaging for Blood Diseases
- Blockchain Technology Applications and Security
Beihang University
2024-2025
Peking University Third Hospital
2013-2024
Peking University
2011-2024
Max Planck Institute for Mathematics
2024
Northeastern University
2024
Huawei Technologies (China)
2022
Kunming University
2022
University of Science and Technology of China
2022
North-West University
2021
Anhui Normal University
2017
Sharding technique is viewed as the most promising solution to improving blockchain scalability. However, implement a sharded blockchain, developers have address two major challenges. The first challenge that ratio of cross-shard transactions (TXs) across shards very high. This issue significantly degrades throughput blockchain. second workloads are largely imbalanced. If imbalanced, some handle an overwhelming number TXs and become congested possibly. Facing these challenges, dilemma it...
Management of spinal neoplasms has relied on open surgery and external beam radiotherapy (EBRT). Although primary tumors are rare, their treatment remains a pervasive problem. This analysis sought to evaluate the safety efficacy CT-guided 125I seed brachytherapy for recurrent paraspinous vertebral tumors. From November 2002 June 2014, 17 patients who met inclusion criteria were retrospectively reviewed. 14 (82.4%) had previously undergone surgery, 15 (88.2%) received conventional EBRT 3...
Abstract Little research has focused on how rotifer communities respond to eutrophication based their combined taxonomic and functional indices. In this research, the relationship of environment was comparatively investigated in two subtropical lakes over one year. The taxon-based indices, including species number ( S ), Margalef index D Simpson d Shannon-wiener H ′), traits relying guild ratio (GR) modified (GR′) from moderately eutrophic Lake Xiyanghu were significantly lower than those...
Hybrid IMRT/VMAT technique which combined intensity modulated radiotherapy (IMRT) and volumetric arc therapy (VMAT) was developed for the treatment of nasopharyngeal cancer (NPC). Two-full-arc VMAT (2ARC-VMAT), 9-field IMRT (9F-IMRT), plans NPC were compared in terms dosimetric quality, sparing organs at risk (OARs), delivery efficiency. The can improve target dose homogeneity conformity with 9F-IMRT 2ARC-VMAT. It reduce delivered to TMJ, mandible, temporal lobe, unspecified tissue fewer MUs...
To develop a 3D-Unet dose prediction model to predict the three-dimensional distribution of volumetric modulated arc therapy (VMAT) for cervical cancer and test performance in endometrial explore feasibility generalization.One hundred seventeen cases 20 treated with VMAT were used training, validation, test. The prescribed was 50.4 Gy 28 fractions. Eight independent channels contoured structures input model, as output model. trained validated on training set (n = 86) validation 11),...
Abstract The incidence of brain metastases is increasing and various treatment modalities exist for metastases. aim this study was to investigate the dosimetric quality delivery efficiency robotic radiosurgery (CyberKnife) multiple compared with C‐arm linear accelerator (linac) based plans. linac plans included intensity‐modulated radiation therapy (IMRT), volumetric modulated arc (VMAT) non‐coplanar VMAT 1, 3 5 arcs, respectively (NC1, NC3 NC5). For 20 patients, six a prescription dose 30...
To explore the implementation of incident learning for quality management radiotherapy in a new established program.With reference to consensus recommendations by American Association Physicist Medicine, an system was specifically reporting, investigating, and individual incidents. The incidents that occurred external beam from February, 2012, 2014, were reported.A total 28 near misses 5 reported. Among them, originated imaging planning, 25 1 plan transfer, commissioning, delivery,...
Based on digital pathology slice scanning technology, artificial intelligence algorithms represented by deep learning have achieved remarkable results in the field of computational pathology. Compared to other medical images, images are more difficult annotate, and thus, there is an extreme lack available datasets for conducting supervised train robust models. In this paper, we propose a self-supervised (SSL) model, global contrast-masked autoencoder (GCMAE), which can encoder ability...
Image perturbation is a promising technique to assess radiomic feature repeatability, but whether it can achieve the same effect as test-retest imaging on model reliability unknown. This study aimed compare based repeatable features determined by two methods using four different classifiers. A 191-patient public breast cancer dataset with 71 scans was used pre-determined 117 training and 74 testing samples. We collected apparent diffusion coefficient images manual tumor segmentations for...
Adversarial example detection, which can be conveniently applied in many scenarios, is important the area of adversarial defense. Unfortunately, existing detection methods suffer from poor generalization performance, because their training process usually relies on examples generated a single known attack and there exists large discrepancy between unseen testing examples. To address this issue, we propose novel method, named Example Detection via Principal Domain Adaptation (AED-PADA)....
Abstract We extend the Hirzebruch–Milnor class of a hypersurface $X$ in an ambient complex algebraic manifold to case where normal bundle is nontrivial and cannot be defined by global function, using associated line graded quotients monodromy filtration. The earlier definition requiring defining function can applied rarely projective hypersurfaces with non-isolated singularities. Indeed, it surprisingly difficult get one-parameter smoothing total space smooth without destroying singularities...
This study aims to evaluate the repeatability of radiomics and dosiomics features via image perturbation patients with cervical cancer. A total 304 cancer planning CT images dose maps were retrospectively included. Random translation, rotation, contour randomization applied before feature extraction. The was assessed using intra-class correlation coefficient (ICC). Pearson (r) adopted quantify between characteristics repeatability. In general, lower compared features, especially after...
Purpose The 3D U-Net deep neural network structure is widely employed for dose prediction in radiotherapy. However, the attention to depth and its impact on accuracy robustness of remains inadequate. Methods 92 cervical cancer patients who underwent Volumetric Modulated Arc Therapy (VMAT) are geometrically augmented investigate effects by training testing three different structures with depths 3, 4, 5. Results For planning target volume (PTV), differences between predicted true values D 98 ,...
生成式人工智能著作权侵权损害难以规制的原因在于其蕴含的“特殊法理”。著作权公益和私益的双重属性孕育了生成式人工智能在侵权主体定位、侵权责任判定路径、损害赔偿承担机制及相当因果关系适用困境的特殊性，进而致使生成式人工智能著作权侵权损害与一般侵权行为存在差异。通过对侵权行为、主体过错、损害事实和因果关系等四方面的比较研究，分层次研析相关法理的具体呈现，从而溯源探析出著作权保护与产业发展、个人收益与公共受益两方面的权益分配不均是著作权侵权损害衍生的本质。结合利益平衡理论，对于生成式人工智能侵权损害的法律规制，应厘清法律保护效能与技术边界范畴，在传统侵权损害规制模式中引入“特殊法理”，建构分层级侵权损害规制模式；依托国家、社会的监管效能，建立涉侵权行为监管体系专项协作配合机制；打造责任主体的侵权豁免机制，营造各利益攸关方的协同治理格局。以此形成兼具公私属性的平衡型保护机制。
Adversarial example detection, which can be conveniently applied in many scenarios, is important the area of adversarial defense. Unfortunately, existing detection methods suffer from poor generalization performance, because their training process usually relies on examples generated a single known attack and there exists large discrepancy between unseen testing examples. To address this issue, we propose novel method, named Example Detection via Principal Domain Adaptation (AED-PADA)....
Generating animatable and editable 3D head avatars is essential for various applications in computer vision graphics. Traditional 3D-aware generative adversarial networks (GANs), often using implicit fields like Neural Radiance Fields (NeRF), achieve photorealistic view-consistent synthesis. However, these methods face limitations deformation flexibility editability, hindering the creation of lifelike easily modifiable heads. We propose a novel approach that enhances editability animation...