Image Processing and Analysis with Graphs: Theory and Practice

http://greyc.stlo.unicaen.fr/lezoray/IPAG/International audienceThe last two decades have witnessed the growth of image production from digital photographs to the medical scans, including satellite images and video films. Consequently the number of applications based on digital images has drastically increased, either for multimedia, computer animation, video games, communication and digital arts, medicine, biometry, etc. Although being very different from one another, all these application areas require similar image processing and analysis techniques. The field of image or video processing analysis is very broad, encompassing a wide variety of research issues from low-level processing (such as image enhancement, restoration, and segmentation) to high-level analysis (semantic object extraction, indexing databases of images, and computer-human interaction). Recently, graphs have emerged as a unified representation for the processing and the analysis of images. The number of concepts that can be defined on graphs is very large. In particular, many real-world problems have been successfully modeled on graphs. Consequently, graph theory has found many developments and applications for image processing and analysis, particularly due to the suitability of graphs to represent any discrete data by modeling neighborhood relationships. Different graph models have been proposed for image analysis, depending on the structures to analyze. However, graphs are not only of interest for representing the data to process, but also for defining graph-theoretical algorithms that enable the processing of functions associated with graphs. This research topic is timely, very influential in computer science and has led to many applications concerning filtering, enhancement, restoration, and object extraction. To this end, image processing and analysis with graphs is way to becoming indispensable for the development of cutting-edge research and applications