Image processing

Expected skills : 32 (322/323), 34 (342/343), 112 (1122)

The image exhibition is a very vast field which has known an important development for some decades and which still know an important activity, as much research level as industrial application level. The applications are numerous, for example the medical imaging (endoscopy, IRM, tomography, echography and so on…), the air or satellite recognitions, the industrial control, the study of the Earth and the atmosphere, the safety and the multimedia.
Generally we distinguish three main categories of operations:
- The image processing: From an input image, one or several output images are generated.
- The image analysis: From one or several input image, measures are calculated to characterize the images or objects presents in these images.
- The image understanding, also named “computer vision”: From one or several input image, we determine a high-level description more or less complex of the scene (for example: nature of present object, topological relations among these objects, reconstituted depths…)
We can establish an inclusive relation among these categories: the image interpretation requires extracting and characterizing of a set of domains, segments or information more or less local (visible movement at the bottom of the image, object or point) by analysis operations of image. They are based on the application of image processing in the aim to extract from the original image the local information necessary to the application (for example: noise decrease, local characterization for the extraction of contours or domains.)
This course is focused on the fundamental concepts of image processing. Some specific concepts to the image analysis will rapidly be introduced (particularly, characterization of the domain). In contrary, the topic of the interpretation (or rather the “high-levels” operators used for the interpretation) should be the object of one or several specific course and won’t be broached.