Project Goal
We wanted to compress any image that is easily broken into 8x8 blocks. We first take
the DCT (Discrete Cosine Transform) of 8x8 blocks of the image, and code each
(first quantizing and then coding the number corresponding to the appropriate
quantization level). We write our bitstream out to a file that can then be
compressed further using gzip or similar commands. We can then read the information
back into Matlab, decode the quantization levels (by using the concatenated
information about which mask was used to compress the DCT) and take the Inverse
DCT to get our original image back (in as good quality as possible-based on the mask used).