Fruit Images Segmentation Based on Fuzzy Art Model

Fruits image segmentation, i.e. classifying the image into homogeneous regions, is a key step of image analysis and computer vision tasks. In this paper, an efficient segmentation method of fruit images by fuzzy clustering is presented. The new approach essentially employs a neural network based on adaptive resonance theory. This approach has the advantage of neural network computation, to improve the precise and robustness of the segmentation. The experimental results have also shown that the proposed method can obtain satisfactory results of fruit image segmentation (to locate stems, to detect blemishes), for the subsequent automatic grading of fruit and image processing systems