AI 728: IMAGE PROCESSING

HANDS ON EXPERIMENTS

These exercises are for gaining hands-on experience with image processing and are open-ended. In other words, there are no off-the-shelf algorithms or techniques that can be borrowed from the web, textbooks, software packages, etc. Please use IPTK and do the problems listed below.

Experiments List

1. Enhance the degraded image given here so that we may be able to read the letter properly. Try to get the best quality possible.
2. De-blur and read the name written on the train engine image. One more image
3. Binarise the document image so that only text is visible on a white page. Extra credit if you write a program that separates the image into three parts: the white paper, Telugu text and the temple image.
4. Identify or separate the face regions in the given image.

Instructions

• The entire class is divided into groups. Please work with your group.
• The first problem is for Groups 1, 5, 9 and 13.
  The second problem is for Groups 2, 6, 10 and 14.
  The third problem is for Groups 3, 7 and 11.
  The fourth problem is for Groups 4, 8 and 12.
• Each group should demonstrate their program and the output on the date given in the course plan.
• As there are at least three groups for any problem, the students (as well as the teacher!) vote for the best result as shown in the demonstration class.
• There will be one winning group for each problem as well as an overall winner
• The winning teams get bragging rights over the others:-)
• As far as coding goes, the programs must be written in IPTK and cannot be copied from each other.
• There are no general rules for which techniques to use, except that they should be related to the material taught in the class until and including the class on 30 Sep 2011.
• GOOD LUCK and HAVE FUN!