A Heuristic Approach to Automatically Segment Signal from Background in DNA Microarray Images

Manjunath, S. S. and Nandihal, P. and Rangarajan, L. (2017) A Heuristic Approach to Automatically Segment Signal from Background in DNA Microarray Images. In: International Conference on Cognitive Computing and Information Processing (CCIP 2017), December 15-16, 2017,, Bangalore.

Full text not available from this repository. (Request a copy)

Abstract

Microarray is an efficacious tool used to detect, analyze and describe local features of the genome in the form of an image. A microarray is glass slide which consists of different nucleic acid inquest added chemically. Microarray provides enough guidance on infection pathology, progression, resistance to treatment, and response to cellular microenvironments and ultimately may lead to improved and innovative approaches for disease diagnosis. Study of microarray includes these steps: initially spot addressing followed by spot separation as background and foreground and finally intensity extraction of separated spots. In the spot separation stage, extended k-means and Fast EM methods are proposed. The proposed method is a heuristic approach where Coefficient-Variation of the input image is used as a metric to switch the image for segmentation to the appropriate existing method as well as to the proposed method to efficiently segment the foreground signals from the background in existence of undetermined or unwanted pixel values, and inadequately articulated spots in DNA microarray images. Experimental results and analysis of proposed methods shows promising results when compared with existing methods in the literature.

Item Type: Conference or Workshop Item (Paper)
Subjects: D Physical Science > Computer Science
Divisions: Department of > Computer Science
Depositing User: MUL SWAPNA user
Date Deposited: 07 Mar 2020 05:05
Last Modified: 07 Mar 2020 05:05
URI: http://eprints.uni-mysore.ac.in/id/eprint/11418

Actions (login required)

View Item View Item