VectorQuantization

LBG (Linde–Buzo–Gray) Vector quantization compression and decompression implementation for gray-scale images. It was an assignment for the Compression and Information theory course in Cairo University Faculty of Computers and Artificial Intelligence.

How to use:

To compress an image you can run the program with the following arguments:

java -jar vectorquantization.jar -c [filename] [vector size] [code book bit count]

For decompression, you can use:

java -jar vectorquantization.jar -d [filename]

Examples:

Using 2x2 vector size and 16 code book size (4 bits)

Input:

Size: 58.9 KB

Output:

Compressed file size: 19.6 KB

Input:

Size: 46.3 KB

Output:

Compressed file size: 19.6 KB