Figure 3From: Application of reinforcement learning for segmentation of transrectal ultrasound images From left to right: original images, ground-truth image (manually segmented), and the results of the proposed approach for eight sample images. Back to article page