Ideology of Image Segmentation and Methods Thereof

Image processing has emerged to be the only stand-alone non-human technique for image comprehension. The avant-garde image processing techniques finds their vast relevance in feature extraction, face detection, finger-print detection, microscope imaging, medical imaging, computer vision, remote sensing, etc. Image processing constitutes a systematic procedure of signal processing, computer vision and artificial intelligence leading to application dependent sorting of unique set of image features like intensity, color, texture, etc. Eventually, segmentation of image has surfaced as a crucial milestone in image processing and its success counts in better results of the trailing steps: be it boundary based (edge detection, watershed, etc.), threshold based (histogram, local, global, etc.) or region based segmentation (region growing, merging, splitting, etc.. This paper intends to explicate a methodical study of a range of image segmentation techniques along with various methods to optimize the results of segmentation. In addition to subjective evaluation, a quantitative survey for the evaluation of the segmented results is also extended.