CLASSIFICATION OF ANALYZABLE METAPHASE IMAGES BY EXTREME LEARNING MACHINES

Authors : Abdülkadir ALBAYRAK

Pages : 78-82

Doi:10.36222/ejt.818160

View : 17 | Download : 9

Publication Date : 2021-06-01

Article Type : Research Paper

Abstract :A chromosome is a DNA molecule that contains the genetic material of an organism. Possible defects in chromosomes can cause structural and functional disorders in living things. Identifying the metaphase stages of cells is a critical step to identify problems in chromosomes. In this proposed study, the discriminative features of possible metaphase images were extracted with Gray Level Co-occurrence Matrix and classified with the Extreme Learning Machines classification method. When the results obtained were evaluated, it was observed that the proposed method was as successful as the deep learning methods in the literature. Especially in recent years, when online learning has become important, the need for re-training of deep learning-based algorithms after each validation will increase the importance of the proposed method in this field. The rapid increase in unlabeled data from each patient every day affects the duration of training and creates time and resource constraints. Fast and accurate modeling of such data with alternative machine learning methods will contribute to the studies in this area.
Keywords : Karyotyping, metaphase detection, extreme learning machines, gray level co occurrence matrix