Image Processing, Pattern Recognition, and Semantic Understanding Techniques

The lengthy first chapter reviews selected methods of initial image analysis, major classification techniques including a list of holistic and syntactic methods, and introduces to techniques of semantic image analysis. This chapter also contains several source codes of example implementations of selected image processing methods. Subsequent paragraphs will discuss in order the three most important groups of methods and algorithms used by modern image analysis systems in more detail. These groups of methods are as follows: Image processing—groups techniques of image preprocessing, which usually allow the quality of the analyzed images to be improved or certain significant objects of interest for further analysis to be extracted from the entire complex image (e.g., during its classification). Pattern classification—a group of methods which allow certain significant features of the entire image or its selected elements to be identified and used for classification, by indexing them as objects which are members of certain previously defined classes. Image understanding—analysis methods aimed at determining the meaning of a given image by analyzing its semantics. Such techniques are very advanced and can be used only for images which contain meaning, e.g., various medical images.