Skip to Main Content

An official website of the United States government

Principal Investigator
Yun Jiang
Peking University
Position Title
Director of Big Data and Machine Learning Innovation Center(MLIC), Associate Dean of EECS
About this CDAS Project
NLST (Learn more about this study)
Project ID
Initial CDAS Request Approval
Sep 28, 2016
Region of interest segmentation in pathology image
Scanning digital images is costly and time consuming as the input pathology image is usually with a high resolution. To mitigate the problem, we propose to segment the region of interest which can be further used for Pathologists to perform diagnoses. Many off-the-shelf algorithms have shown promising results on objects segmentation (e.g. dog, person, horse). So it seems to be interesting to build a segmentation based on these models. From our perspective, there are two main difference between region of interest segmentation in pathology image and object segmentation in natural images. Firstly, the valid data is rare for region of interest segmentation in pathology image. Therefore, it is very hard to train a deep network to perform well. Secondly, the resolution between these two kinds of image is various. Therefore, we need to develop a novel network structure or optimization algorithm which is appropriate for pathology images.