Vector Quantization Here are the steps that you need to implement to compress an image.
- Understanding your two pixel vector space to see what vectors your image contains
- Initialization of codewords - select N initial codewords
- Clustering vectors around each code word
- Refine and Update your code words depending on outcome of 3. Repeat steps 3 and 4 until code words don’t change or the change is very minimal.
- Quantize input vectors to produce output image