UTHM Institutional Repository

Abnormal gastric cell segmentation based on shape using morphological operations

Abdul Khalid, Noor Elaiza and Samsudin, Nurnabilah and Hashim, Rathiah (2012) Abnormal gastric cell segmentation based on shape using morphological operations. In: ICCSA'12: Proceedings of the 12th International Conference on Computational Science and Its Applications, 18-21 June 2012, Salvador de Bahia, Brazil. (Unpublished)

Full text not available from this repository.

Abstract

Cancer is the fourth leading cause of death among medically certified deaths in Malaysia. The most reliable diagnostic method to diagnose gastric adenocarcinoma is by inspecting the microscopic images of samples obtained through biopsy. These images are analyses by pathologist to identify the presence of cancer. However the process is time consuming and the interpretation varies with different pathologist. The application of image analysis techniques can assist pathologist towards a more efficient and faster diagnosis. Thus, this paper introduces an image analysis framework to automatically recognize and distinguished between normal gastric and gastric adenocarcinoma cells. The framework consist of the three phases of image analysis; preprocessing phase where the color tone issues are solved by component separation; processing phase which includes the thresholding and morphological techniques to segment the cells; post processing to identify the perimeter, area and roundness of the cells. This study shows that it is possible to automatically recognize and differentiate images with normal and abnormal cells.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: abnormal cell detection; image processing; morphological operation; roundness; segmentation; stomach
Subjects: R Medicine > RZ Other systems of medicine
Depositing User: Normajihan Abd. Rahman
Date Deposited: 13 Aug 2018 03:35
Last Modified: 13 Aug 2018 03:35
URI: http://eprints.uthm.edu.my/id/eprint/3737
Statistic Details: View Download Statistic

Actions (login required)

View Item View Item