Review of Image Analysis
Sandip Dey, Siddhartha Bhattacharyya, Ujjwal Maulik · 2019
This chapter comprises a brief discussion of the review of image analysis. The objective of this technique is to introduce conventional and other new advanced versions of digital image processing methodologies that primarily engage the image analysis class of tasks, namely, image segmentation, feature description (extraction and selection), and object classification. The first stage of image analysis techniques target applying different image segmentation techniques. The most common methodologies in this direction involve thresholding, region-based segmentation, boundary-based segmentation, and texture segmentation. The chapter presents the mathematical formalism of image segmentation technique. Image thresholding is a well-known, simple, and frequently used technique in image segmentation. Using thresholding as a tool is effortless to accomplish bi-level and multi-level image thresholding and it provides robust results in all conditions. Image thresholding appears to be an optimization problem where plenty of objectives like Otsu's function and Kapur's function can be used as the optimization function.