Thursday, August 27, 2020 - 13:00 to 14:00
SCHOOL OF COMPUTER SCIENCE
The School of Computer Science is pleased to present…
MSc Thesis Proposal by: Sameer Akhtar Syed
Date: Thursday August 27, 2020
Time: 1:00 PM – 2:00 PM
Zoom URL: https://zoom.us/j/91480137746?
Cancer is an important public health problem and the third most important cause of death in North America. The American Cancer Society (ACS) estimates that by the end of 2020, 1,806,590 new cases and 606,520 deaths will occur in the US. Among the highest impact types of cancer are colorectal, breast, lung, and prostate. Different algorithms provide different solutions for CAD purpose. For example, classification algorithms are employed in identifying histological structures, such as lymphocytes, cancer-cells nuclei and glands, from features like color, texture or shape. The role of color characteristics in these images is still an issue that needs to be investigated. This research focuses on the relationship between color and morphological characteristics in a dataset consisting of 5000 histopathological images in CRC histology with 8 classes. The obtention of the color characteristics will be done with first-level statistics, which are the mean, median, standard deviation, kurtosis, skewness and the color histogram. Morphological features are extracted using features like circularity, eccentricity, solidity, area, and perimeter. Finally, four machine learning algorithms are applied and their performance is measured on Accuracy, Precision, Recall and F1- Score. When compared with other works, the performance of our method stands in the same range.
Keywords: Automated Diagnosis, Colorectal Cancer, Computer Vision,
Internal Reader: Dr. Imran Ahmad
External Reader: Dr. Mohamed Belalia
Advisor: Dr. Boubakeur Boufama