Question from a Fisheries Scientist [R]

So, I count fish from side scan sonar images and add up the totals. The fish pretty much just look like small oval bright spots with shadows near them. For the past year I have been experimenting with quicker ways to count them like imagej/Fiji, Weka, and some specific image segmentation python libraries. Recently I have been trying to look into deeplearning architectures for counting my fish and have been also experimenting with Mask R-CNN for object detection and instance …

count fish image images libraries look machinelearning near python question segmentation small sonar them weka