Prostate Gland Segmentation in Histology Images via Residual and Multi-Resolution U-Net
This work addresses prostate cancer diagnosis by improving gland segmentation accuracy for computer-aided systems, though it appears incremental in method.
The paper tackled prostate gland segmentation in histology images by modifying U-Net architectures with residual and multi-resolution blocks, achieving an average Dice Index of 0.77 and outperforming previous state-of-the-art approaches.
Prostate cancer is one of the most prevalent cancers worldwide. One of the key factors in reducing its mortality is based on early detection. The computer-aided diagnosis systems for this task are based on the glandular structural analysis in histology images. Hence, accurate gland detection and segmentation is crucial for a successful prediction. The methodological basis of this work is a prostate gland segmentation based on U-Net convolutional neural network architectures modified with residual and multi-resolution blocks, trained using data augmentation techniques. The residual configuration outperforms in the test subset the previous state-of-the-art approaches in an image-level comparison, reaching an average Dice Index of 0.77.