Incorporating The Contour Information Into The Region-Based Image Segmentation Algorithm

In today’s society of the rapid development of science and technology, an important part of the science and technology—computer science and technology, should not be ignored, and its application have been we unconsciously infiltration into the nook and cranny of our society. In the practical application of daily life, we more and more contact and apply the computer vision, the unmanned aerial vehicle technology, the skin care software and so on. So we can’t ignore the many image processing problems in the real world. Especially the unmanned aerial vehicle and the intelligent robots having artificial intelligence machines need to proceed certain image recognition processing work independently. Image processing involves a variety of processing algorithms and works, but the application can’t leave the basic, core image segmentation algorithm. Therefore, in the research field in image processing, image segmentation problem is still a fundamental problem, and is the current hot spot in the field of image processing.At present, image segmentation problem is still one of the important research topic in the field of image processing. The current due to the rapid development of practical application, the image processing problem for the natural images is more and more get people’s attention. Due to the complexity and variability of the natural images, image processing problem for the natural images still has a lot of research on space. Usually, in the field of the visual processing, the result of image segmentation is as the basis step of various field of computer image processing application. Therefore, image segmentation algorithms can be thought of as the key step in the various image processing applications. But because of the complexity of the natural images, such as a natural image will involve multiple objects, and the influence of light and shadow is very large. For the complexity of the natural images, at present most of the segmentation results of the segmentation algorithms are not very ideal, and are not applicable to complicated natural images.New method is put forward in this paper the Boundary - MLAP image segmentation algorithm, with the traditional image segmentation algorithm, has its own advantages. In a nutshell, the traditional image segmentation algorithms are mainly divided into two categories:one is image segmentation method based on region, another kind is image segmentation method based on contour. And the new image segmentation method is different, this article puts forward a new method of the outline of the image information into image segmentation algorithm based on region, realize the image segmentation method based on region and the combination of the image segmentation method based on contour, which can take advantage of the image contour information advantages and the advantages of image segmentation algorithm based on region, combination of these two together improve the quality of natural image segmentation.The Boundary-MLAP image segmentation method is based on the MLAP image segmentation framework. The main characteristic of the MLAP image segmentation algorithm framework can use the variety image features which are extracted from the segmentation image itself for image segmentation. Based on the MLAP image segmentation framework, the new image segmentation algorithm the Boundary-MLAP do the algorithm research work from two aspects:on the one hand, to optimize the original MLAP image segmentation algorithm framework, and make full use of the respective advantages of the variety image characteristics to segment the image; On the other hand, incorporate the contour information which is extracted from itself into image segmentation algorithm of optimized framework, so as to realize the image segmentation algorithm based on region and the combination of the image segmentation algorithm based on contour, the implementation of the final Boundary-MLAP image segmentation algorithm is proposed. In this paper, the proposed image segmentation method realize to image contour information optimization in the image segmentation algorithm based on region to make the result of natural color image segmentation is more reasonable. The image segmentation results contrast in the subsequent parts of the article, by comparing the different image segmentation algorithms which are natural experimental results of image segmentation, fully proved, in the treatment of natural images, in the segmentation of natural images results, the new image segmentation results got by the new image segmentation algorithm proposed in this paper are more reasonable, and the contour images got by the segmentation results are more figure closer to the real contour, especially on the various categories of natural images segmentation, it has more advantages. In the comparison of experimental results, this paper gives the detailed numerical parameters and the compare results.