Abstract:

The fish fecundity in an important parameter tomanage sustainable fisheries. Traditionally, its calculation ismanually performed on the histological images of fish ovaries,counting and measuring its matured reproductive cells, which is avery time consuming process. The automatization of this processimplies the recognition of the matured cells in the image. Thispaper compares the statistical performance of five state-of-artsegmentation techniques, which code is publically available, tosegment histological images of fish ovary in order to recognizethe outline of its cells. The approaches based on Canny filter andK-means clustering provides the best trade-off in performanceand speed for both fish species tested. Although other approachesprovide comparable performances at pixel level, we are interestedin their efficiency at region level (the recognition of the outlineof the cells) in order to measure them and calculate the fishfecundity