Image segmentation based on multi-scale local feature

Image segmentation is the basic problem of image processing and computer vision, and it´s also the key step from image processing to image analysis. The existing image segmentation methods were firstly reviewed and then the concept of multi-scale local features was introduced. A region-driven image segmentation of multi-scale local energy was proposed later. Finally, these algorithms were evaluated by the manual segmentation result and the evaluation results showed that the method based on multi-scale local energy could obtain accurate segmentation results with high availability.