Image processing

Nowadays, image processing is omnipresent in information technologies: medicine, biology, agriculture, entertainment, culture, measurement, mechanics...

Image processing consists of applying mathematical transformations to images in order to modify their appearance or extract information from them. More generally, image processing aims to manipulate the underlying information contained in an image. Although it has long been carried out using electronic circuits, image processing is nowadays carried out almost exclusively digitally, i.e. via algorithms generally programmed with an imperative language (C, C++, Java, Python, etc.).

This teaching unit aims to give a solid foundation in image processing. It addresses, among other things, image formation and acquisition, colorimetric transformations, morphological operations, geometric transformations, compression, frequency transformations, recognition and matching techniques, etc. and an introduction to deep learning methods. The courses are complemented by support videos.

The teaching unit is mainly composed of 11 didactic courses covering the basics in the main areas of image processing and 3 practical work sessions whose subjects are to be chosen from 6 proposals. Students can choose to carry out the work on images they bring corresponding to their field of training.

The aim of this module is to provide students with a basic understanding of image processing, enabling them 1/ to understand the operations performed by image processing software, 2/ to read articles on image processing, 3/ to develop their own applications and 4/ to pursue their own training in this field.

Signal processing basics.

Some programming skills in a language.

Written 60%, Practical work (based on report) 40%.

• Image formation, digital image, luminance image, color image, ...
• Image acquisition technology.
• Mathematical morphology
• Convolution kernels, interpolation kernels: discrete representation of transformations defined in the continuous domain.
• Image derivation.
• Extraction of contours and special points.
• Fourier transform on images.
• Image filtering, digital convolution and noise reduction.
• Correlation and distances between images.
• Principle of image compression.
• Geometric transformations.

CM: 4:30 p.m.

 Practical work: 9:00 a.m.