A Comparison of YOLO and Mask-RCNN for Detecting Cells from Microfluidic Images

Abstract

As an effective model to study aging, the budding yeast Saccharomyces cerevisiae has revealed aging mechanisms that are shared with human aging. Yeast cell lifespan can be measured in replicative lifespans (RLS) - the number of cell divisions from a single mother cell before dying. However, counting yeast cell divisions from microscopic images is a tedious task. Here, we address this challenge with computer vision object detection. We compared two deep learning methods, YOLO and MASK R-CNN to detect cells from microfluidic images. We concluded that YOLO is more sensitive at detecting cells, whereas MASK-RCNN is more informative on cell sizes. Therefore, both methods are useful for automatic microfluidic image analysis.