Advanced search
Open CourseWare >> Mathematics and Applications >> Applied Mathematics and Computing
Teaching coordinator :

Beatriz Marcotegui
  

Teaching staff :
Fernand Meyer
Serge Beucher

Research center

Level : Graduate

Course Language : English

Term : Fall

Number of hours : 35

ECTS Credits : 0
IC-OTA1 Advanced Morphological Segmentation
Teaching site :
The Centre for Mathematical Morphology (CMM), specialized in image processing, is a common research centre to ARMINES and the Ecole des Mines de Paris.The CMM has created a theory and a method for image processing, called mathematical morphology. CMM has acquired a leading position in the fields of mathematical foundations of the theory, efficient algorithms design, resolution of applications involving image processing in the bio-medical, material and multimedia fields.

Program
The watershed is the basic tool of Mathematical Morphology for segmentation. It has proved to be powerful and it is used in a large number of applications, such as, medicine, remote sensing, robotics, multimedia and character recognition. This algorithm is the starting point of the course and then more recent developments are presented. This course gives a comprehensive presentation of the morphological approach to image segmentation, supervised or unsupervised, in 2D or 3D, for grey-tone or colour images, for still images or sequences as well as related techniques necessary for successful applications. The program of the course is:

* The watershed : the use of markers to deal with over-segmentation problems.
* Pre-processing techniques : we present a class of morphological filters called levelings, able to simplify the images without blurring the contours, and without creating regional minima or maxima. They are the basis of a morphological scale-space representation of images.
* Multi-scale segmentation : a multi-scale segmentation approach, combined with the graph theory, is presented. It leads to efficient algorithms and synthetic image representations for image interpretation. Strategies for interactive segmentation.
* Segmentation of colour images.
* Segmentation in 3D. Segmentation of sequences.
* Presentation of different applications.

Structure of the course
Five full days in a single week. Lectures in the morning and practical training on PC computers in the afternoon. Demos.

The week before the course, three additional days of course usually take place. They are devoted to an Introduction to Mathematical Morphology.

Requirements : This course is built on the Mathematical Morphology theory.

Last Modification : Thursday 26 July 2007

© Mines de Paris 2019 - Developed by Winch Communication