Analysis Of Image Segmentation Techniques

Image segmentation aims to partition an image into segments of homogeneousproperties. It is a fundamental step in various image-related computation and analysis. Itis the basis of further studies such as region extraction and pattern recognition. To meetthe needs of modern issues such as face and fingerprint detection, medical imagerecognition, and other advanced applications, a wide range of novel and efficient imagesegmentation algorithms have been proposed these years. In this work, a newsegmentation technique making use of connected component analysis is designed,which aims to improve the efficiency of the segmentation process and the quality of thesegmentation results. Test and evaluation of some well-known segmentation toolboxesand systems employing different algorithms are also performed in this project.Comparison of the proposed segmentation system with those tools reveals that the newsegmentation method is a good trade-off between complexity and efficiency.