This project comes from the multimedia signal processing course at National Taiwan University of Science and Technology. This project contains four themes :

• Neighborhood Process - Error Diffusion
• Huffman Coding
• Point Process-Ordered Dithering
• Vector Quantization
• 2D Discrete Cosine Transform

This image is original ena as a control group.

Using the Classical-4 & Bayer-5 Dither Array.

Classical-4

Bayer-5

Using the Floyd_Steinberg, Jarvis and Stucki.

Floyd_Steinberg

Jarvis

Stucki

Huffman code is a particular type of optimal prefix code that is commonly used for lossless data compression. There is the example image to show how huffman coding work.

Example Image

This project includes the implementation of Vector Quantization(VQ). The block size are 4*4 and implement with for 4 different codebook size image are shown below:

Codebook Size = 16

Codebook Size = 64

Codebook Size = 256

Codebook Size = 1024

Including 3 different block size with Spatial Domain image, Frequency Domain image and DCT.