The aim of this module is to acquire a basic understanding of computer algorithms for image interpretation and to appreciate the strengths and limitations of digital imagery.

In this module we will show the complexity and challenges of image analysis and understanding. We will see why knowledge in mathematics (e.g. linear algebra), programming (e.g. OpenCV libraries in C++ or Matlab), computer vision, etc. are necessary to select the best algorithm to analysis a complex image.

First, we will introduce the basics of image processing & analysis (such as image registration, segmentation, representation and modeling), we will show from case studies how difficult image processing & analysis actually is. Next, we will introduce the basics of image understanding (such as object detection, recognition and tracking, scene analysis). We will show from examples what approaches and heuristics can be used to solve practical problems such as: – geometric feature extraction, – low level features extraction, – multi-resolution and multi-scale analysis. At the end, we will see how machine learning (such as neural network techniques) and data analysis approaches (such as clustering and classification techniques) have pushed back the limits of current image analysis models and also open new perspectives in this domain.

Lastly, a special focus on most classical methods will be made on simple examples using C++ or Matlab, and more specially the Image Processing Toolbox. Then each student at the end will be able to manipulate images and extract their intrinsic content for real applications.