The goal of this project is to develop stochastic image analysis and computer vision techniques to obtain accurate volume computation of tumor. Due to complex structures of different normal tissues such the gray matter (GM), white matter (WM), and cerebrospinal fluid (CSF) and abnormal tissues such as cyst, edema, necrosis and scar in the MR brain images, robust measurement of tumor is challenging. We have developed multi-resolution fractal stochastic feature extraction, unsupervised clustering and information theoretic approach to obtain robust tumor segmentation. Currently we have extended our method to segment the multiple brain abnormal tissues (edema, necrosis, enhanced tumor core and non-enhanced core) and ischemic stroke lesions. Participation in multiple Global tissue Segmentation Challenges provided state-of-art comparisons for our tumor and lesion segmentation methods. Our methods ranked 3rd out of 8 in BRATS-2013[1], 4th out of 15 in BRATS-2014[2] and 4th out of 14 in ISLES-2015[3].

[1] (Brain Tissue Segmentation Challenge (BRATS) – NCI and MICCAI, Japan)
[2] (BRATS Challenge – NCI and MICCAI, Boston)
[3] (Stroke lesion segmentation Challenge- MICCAI, Germany)


Fig 1: Input images & segmented results for tumor (Top) and lesion (Bottom). Col (1-4): Input images, Col 5: Segmented results, Col 6: Expert’s outline.

The ODU Automatic Tumor Detection Tool has been developed in order to provide a solution to the detection of brain tumors. This tool is made available for evaluation. In order to obtain the tool, please see the external resources page and complete the form, and we will provide you with a download link. To learn more about the tool, please see the video demonstration below. Please be sure to read the documentation before using the tool.

This project is partially sponsored by and/or in collaboration with the following: NIH, CHOP