Research Of Interactive Image Segmentation On Mobile Phone

Image segmentation, an important part of the image processing, has made plenty ofsignificant research after a rapid development from the blue screen split to thesegmentation of the complex background image. The development of Image segmentationnot only has brought convenience to people’s lives, but also has been used to videoproduction, which provides a splendid visual experience to people as well as greatcommercial profit to the film maker. In addition, image segmentation has become moreand more available on mobile devices with the development of hardware technology,especially the improvement of the data processing capacity in mobile devices. Interactiveimage segmentation methods make the results of image segmentation meet users’expectations more by combining users’ subjective willing and objective recognition withfast data processing capabilities of computers or mobile devices.This paper introduces the basic principle and methods of the non-interactive andinteractive image segmentation technology in digital image processing. We completed theGrab Cut algorithm on PC and made a comparison and analysis to its experimental results.In addition, the Grab Cut algorithm has been applied to the interactive image segmentationsystem on the Android platform of mobile phone. By analyzing its experimental results,the factors that affect the efficiency of the algorithm have been put forward, preparing forthe further improvement of the algorithm.Firstly, this paper completed the interactive image segmentation system on PC andAndroid platform respectively. Secondly, their experimental results show that factorswhich have great influences on the results of the image segmentation are the resolution ofthe target image, the number of iterations and the interactive operation of the user.Studying the impact of these factors on the image segmentation’s results also need to bedone in the future, which has great significance on the improvement of the existing imagesegmentation algorithms. Thirdly, this paper proved that the CPU load capacity of phonecan meet the accuracy demand of image segmentation, which needed to be done in thefuture in order to meet the real-time demand.