I begin to develop an algorithm, which has the purpose to identify the area of the leaf that contains spots to report the severity of a disease. Therefore, during my research, I found out about LeafSnap and tried to follow their paper to segment the leaf on the image using OpenCV Expectation Maximization, which is trained using S and V form HSV color space; however, it still return me some false positives or part of the image that are not pointed out to a cluster.
Which techniques can I apply to get some better results?
