A study of machine learning techniques for Automated Karyotyping System

Genetic abnormalities constitute a considerable share of all the existing societal healthcare issues. There has been a dire need for the automation of chromosomal analysis, hence supporting laboratory workers in effective classification and identifying such abnormalities. Nevertheless, with many modern image processing techniques, like Karyotyping, improved the life expectancy and the quality of life of such cases. The standard image-based analysis procedures include Pre-processing, Segmentation, Feature extraction, and Classification of images. When explicitly considering Karyotyping, the processes of Segmentation and Classification of chromosomes have been the most complex, with much existing literature focusing on the same. Various model-based machine learning models have proven to be highly effective in solving existing issues and building an artificial intelligence-based, autonomous-centric karyotyping system. An autonomous Karyotyping System will connect the pre-processing, Segmentation, and classification of metaphase images. The review focuses on machine learning-based algorithms for efficient classification accuracy. The study has the sole motive of moving towards an effective classification method for karyotype metaphase images, which will eventually predict the fetuss abnormalities more effectively. The studys results shall benefit future researchers working in this area.