Colony image classification on fuzzy mathematics

Abstract

Due to colony image quality variation very much, an ordinary colony delineation algorithm is difficult to segment all kinds of colony images, therefore, image classification is necessary before image segmentation. The developed special colony image classification method in this study is to use definition of Judgment Set, determination of Fuzzy Judgment Matrix, and defining weight Set based on colony image characteristics, which are: (1) colony density; (2) colony area percentage (the ratio between colony area and whole area of the image); (3) colony area variance; and (4) grey contrast between colony and nutrient fluid. Experiments prove that the studied method make the classification reasonable, it can be used for colony image recognition and image pre-segmentation, and can also be expanded into the other similar applications.