Weakly supervised semantic segmentation of tomographic images in the diagnosis of stroke
This addresses the challenge of automating stroke diagnosis from medical images, but it is incremental as it builds on existing U-Net architectures with modifications for handling label noise.
The paper tackles the problem of segmenting acute stroke areas in non-contrast CT brain images using a weakly supervised approach, where some training data has inaccurate labels due to radiologist errors, and reports that the proposed methods increase segmentation accuracy.
This paper presents an automatic algorithm for the segmentation of areas affected by an acute stroke on the non-contrast computed tomography brain images. The proposed algorithm is designed for learning in a weakly supervised scenario when some images are labeled accurately, and some images are labeled inaccurately. Wrong labels appear as a result of inaccuracy made by a radiologist in the process of manual annotation of computed tomography images. We propose methods for solving the segmentation problem in the case of inaccurately labeled training data. We use the U-Net neural network architecture with several modifications. Experiments on real computed tomography scans show that the proposed methods increase the segmentation accuracy.