Get our free extension to see links to code for papers anywhere online!

Chrome logo Add to Chrome

Firefox logo Add to Firefox

Constrained Deep Weak Supervision for Histopathology Image Segmentation

Jan 03, 2017
Zhipeng Jia, Xingyi Huang, Eric I-Chao Chang, Yan Xu

Share this with someone who'll enjoy it:

In this paper, we develop a new weakly-supervised learning algorithm to learn to segment cancerous regions in histopathology images. Our work is under a multiple instance learning framework (MIL) with a new formulation, deep weak supervision (DWS); we also propose an effective way to introduce constraints to our neural networks to assist the learning process. The contributions of our algorithm are threefold: (1) We build an end-to-end learning system that segments cancerous regions with fully convolutional networks (FCN) in which image-to-image weakly-supervised learning is performed. (2) We develop a deep week supervision formulation to exploit multi-scale learning under weak supervision within fully convolutional networks. (3) Constraints about positive instances are introduced in our approach to effectively explore additional weakly-supervised information that is easy to obtain and enjoys a significant boost to the learning process. The proposed algorithm, abbreviated as DWS-MIL, is easy to implement and can be trained efficiently. Our system demonstrates state-of-the-art results on large-scale histopathology image datasets and can be applied to various applications in medical imaging beyond histopathology images such as MRI, CT, and ultrasound images.

   Access Paper Source

Share this with someone who'll enjoy it: