基于动态多实例学习的病理图像分类算法研究
dc.contributor.advisor | 黄绍辉 | |
dc.contributor.author | 吴炜 | |
dc.date.accessioned | 2024-05-08T01:17:50Z | |
dc.date.available | 2024-05-08T01:17:50Z | |
dc.date.issued | 2023-06-08 | |
dc.date.replied | 2023-05-16 | |
dc.description.abstract | 病理图像是癌症诊断的金标准,对病理图像的分析有助于医生更准确地诊断癌症类型、病程和预后,并制定更有效的治疗方案。然而,由于病理图像通常具有极高的分辨率以及复杂的结构,增加了对病理图像自动分析的难度。随着数字病理学的发展,基于深度学习的病理图像分析方法取得了一定的成果,但仍然存在诸多挑战。本文从“动态”的角度出发,提出了三种应用于病理图像分类任务的动态多实例学习方法: (1) 基于聚类的自适应多实例学习方法。该方法首先利用聚类算法将病理图像块分组并构成表型聚类集合,动态特征选择器能够自适应地选择具有判别性的特征集合,同时生成表型聚类对应的表型级别特征。该方法能够将图像块特征数量平均减少85%以... | |
dc.description.abstract | Pathological image is the gold standard for cancer diagnosis, and the analysis of pathological images can help doctors to diagnose cancer types, progression, prognosis more accurately, and formulate more effective treatment plans. However, due to their extremely high resolution and complex structure, the analysis of pathological images in­ creases the difficulty of automatic analysis. With th... | |
dc.description.note | 学位:工学硕士 | |
dc.description.note | 院系专业:信息学院_计算机科学与技术 | |
dc.identifier.uri | https://dspace.xmu.edu.cn/handle/2288/217895 | |
dc.language.iso | zh_CN | |
dc.source.uri | https://etd.xmu.edu.cn/detail.asp?serial=817FC2E1-6780-4DBA-98BB-65A1DA83E107 | |
dc.title | 基于动态多实例学习的病理图像分类算法研究 | |
dc.title.alternative | A Study on Pathological Image Classification Algorithm Based on Dynamic Multi-instance Learning | |
dc.type | thesis |
Files
Original bundle
1 - 1 of 1