Adaptive deep learning for visualisation of diagnostic features of dermoscopy images

自适应深度学习用于皮肤镜图像诊断特征的可视化

• 批准号: 490549-2015
• 负责人: Mori, Gregory
• 金额: $ 1.82万
• 依托单位: Simon Fraser University
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2015
• 资助国家: 加拿大
• 起止时间: 2015-01-01 至 2016-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=586521
• 关键词: Adaptive; deep; learning; visualisation; diagnostic

Skin cancer is the dominant form of cancer in North America. Fortunately, it is also the most treatable if detected early. In the clinical setting, the screening procedure typically consists of direct visual examination of the skin, taking digital photographs of suspicious skin lesions with a specialized tool (a dermoscope), and acquiring biopsy samples. However, the wait time for a dermatologist appointment can be up to several months, and in remote areas access to medical facilities can be difficult. With the emergence of smartphones incorporating high-resolution cameras, it is now possible to perform personalized skin imaging at home with a specialized low-cost optical smartphone attachment, and images can be transferred to a doctor for diagnosis. During the collaborative project between MetaOptima and the SFU research team, we wish to improve the accuracy of MetaOptima's algorithms for image analysis and search-by-similarity.

皮肤癌是北美的主要癌症形式。幸运的是，它也是最可治疗的，如果 很早就发现了。在临床环境中，筛查程序通常包括直接目视检查 皮肤，使用专门工具(皮肤镜)拍摄可疑皮肤损害的数字照片，以及 获取活组织检查样本。然而，皮肤科医生预约的等待时间可能长达几个小时。 几个月的时间，在偏远地区，获得医疗设施可能很困难。随着智能手机的出现 结合高分辨率摄像头，现在可以在家中使用 专门的低成本光学智能手机附件，图像可以传输给医生进行诊断。 在MetaOptima和SFU研究团队的合作项目中，我们希望改善 MetaOptima用于图像分析和按相似度搜索的算法的准确性。