Computer Vision News 18 Challenges in Medical AI Development Lacking sufficient data in medical AI is critical because of the vast number of acquisition devices: even simple things like CT segmentation of the anatomy (for instance, airways segmentation or bone segmentation), not to mention more complex tasks such as pathology detection, requires significant variability. Data is key in all AI projects. Training and optimization of neural networks rely on large quantities of data. This is all the more critical for medical AI. It is also true that medical AI projects often lack a sufficient amount of data. The deficiency might be in quantity and/or in quality. When developing a next-generation medical device, data may even be totally absent. Let’s discover some mitigation tips by RSIP Vision (part 1) Ilya Kovler, CTO at RSIP Vision
RkJQdWJsaXNoZXIy NTc3NzU=