Improving Cancer Research with Computer Vision

Computer vision is significantly enhancing life sciences by enabling more accurate diagnoses and research methods, particularly in cancer research. Mateo Sokac, a researcher at Aarhus University and Aarhus Universitetshospital, is utilizing computer vision to improve the detection of neutrophils, a vital type of white blood cell, through object detection. By training models on images from immersion and dry microscopy, Sokac aims to make the identification of neutrophils as accurate in dry microscopy as it is in immersion. His team employs image tiling to manage large input images, which allows for high-quality model training and inference, leading to improved results with a mean average precision increase of 10 percent. This technique not only accelerates experimental processes but also enhances the accuracy of results, thereby advancing cancer research.