An Efficient Mean Shift and Graph Based Image Segmentation
P. Kavitha, Sandhya Prabhakaran · 2014
For some applications, such as image recognition or compression, we cannot process the whole image directly for the reason that it is inefficient and unpractical. Therefore, several image segmentation algorithms were proposed to segment an im-age before recognition or compression. Image segmentation is to classify or cluster an image into several parts (regions) according to the feature of image, for example, the pixel value or the frequency response. Up to now, lots of image segmentation algo-rithms exist and be extensively applied in science and daily life. According to their segmentation method, we can approximately categorize them into region-based seg-mentation, data clustering, and edge-base segmentation. In this tutorial, we survey several popular image segmentation algorithms, discuss their specialties, and show their segmentation results. Moreover, some segmentation applications are described in the end.