AN ARTIFICIAL INTELLIGENCE APPROACH TO DETECT AND CLASSIFY DNA INTEGRITY OF SPERMS AND EGGS USING YOLO ALGORITHM

Abstract

Abstract: This research focuses on proposing the use of YOLO (You Only Look Once) algorithm for detecting and categorizing DNA health condition of sperms and eggs, which is very significant in reproductive health. The data set consisted of 10 000 images of sperms and eggs that have been labeled, and were split into training, validation, and test datasets. The YOLO algorithm showed very high accuracy of 97%. 5% better than previous techniques in identification of DNA irregularities while at the same time being 3% faster. Comparing YOLO with such algorithms as Faster R-CNN and SSD, we can see that the precision rate of this algorithm grows into the mark of 96. Recall rate recorded was 95.3% while on yield rate only 3% responded to the study. 8 %, which is better for real time applications. The findings of the study conveys the benefits of the application of AI and deep learning in improving diagnostic precision and speed in Reproductive medicine. Based on the results, people get new opportunities of studying AI applications in healthcare diagnosis that aim at developing more profound approaches to the issue.