Quantifying clutter in images

• Main Author: Lim, Boon Chieh (Lin Wen Jie)
• Other Authors: Deepu Rajan
• Format: Final Year Project
• Language: English
• Published: 2015
• Subjects: DRNTU::Engineering::Electrical and electronic engineering
• Online Access: http://hdl.handle.net/10356/62828
• Institution: Nanyang Technological University

An image is made up of pixels. Pixels are dots with its own information or color. The number of such pixels are dependent on the dimensions and the fewer the pixels, the less information it can contain. A usual pixel has 24 bits, made up by 3 colors of 8 bits each, which is why a lot of information can be hidden in it. This project however does not use the usual RGB 24-bit color image. The normal image is converted into its CieLAB format, compromising of values for luminosity, and chrominance A and B, the blue (and green) and the red (and yellow) components of the image. Shannon’s entropy is the measure of uncertainty in an image rather than its certainty. This project uses Shannon’s entropy to measure amount of information in an image and then uses probability distribution to collate a value with respect to the information. The codes in the project will give values to each pixel so that it can be quantized. By the end of the project, each image that uses this code to run will have a value associated with it. This value represents the visual clutter value calculated from the breakdown of luminance, and chrominance. Future improvements can either be using a different programming language to measure the same visual clutter value, or devise another method to measure it. MATLAB is not a commonly taught and learnt programming language. A more familiar programming language is advised to be used instead.