Computer Vision

Principles of digital processing of medical images (pixel, point operators, convolution, non-linear filters). Image and video Segmentation (active contours, mean shift, graph-based techniques) Partial Differential Equations and image processing (eg. diffusion, –) 2D motion (optic flow, object tracking -Lukas-Kanade, Meanshift, Kalman) Parameter estimation (μετασχ. Hough, μέθοδος ελαχίστων τετραγώνων, RANSAC, active shape models) Geometric tansformations (affine, projective, local-elastic), image registration 3D visualization (3D to surface/volume rendering, image fusion) Image descriptors (texture, area and border descriptors, local image structure: Hessian and Jacobian matrix), Salient points in images (Harris, SHIFT, SURF), Pattern recognition techniques Camera calibration (pinhole model, special wide angle/omnidirectional cameras). Applications: 3D clues from mono-occular and bi-occular images, shape from silhouettes Prerequisites The students should be familiar with basic concepts of Calculus, Linear Algebra, Numerical Analysis and Matlab programming

• Comprehension of the main concepts of Computer Vision.
• Learning the mathematical formalism and relevant algorithms.
• Acquiring abilities to implement algorithms for processing and analysis of biomedical images.
• Analyzing relevant problems and selection of the proper filters/operators. Understanding the requirements specific to processing and analysis of biomedical images.

• Digital Image Processing, Gonzales, “Εκδόσεις Α. ΤΖΙΟΛΑ & ΥΙΟΙ Α.Ε.”, 3η έκδ./2010 ΘΕΣ/ΝΙΚΗ, Eudoxus code: 18548692
• Ψηφιακή Επεξεργασία και Ανάλυση Εικόνας, Παπαμάρκος Νικόλαος, ΝΙΚΟΛΑΟΣ ΠΑΠΑΜΑΡΚΟΥ, Εκδ. 2η/2010, ΑΘΗΝΑ
• Ψηφιακή Επεξεργασία Εικόνας, Ι, Πήτας, Ιωάννης Πήτας, Εκδ. 2η/2010, ΘΕΣ/ΝΙΚΗ, 8020