Our work at a glance
We have primarily implemented the JPEG baseline compression method for gray
scale images. The steps followed are as follows:
- Shift the levels of the pixel values by 128 downwards
- Get 8x8 DCT for the whole image.
- Quantize the image using standard JPEG quantization matrices.
- Order the the quantize image in Zig-Zag manner so as to enable run-length variable
length encoding.
Salient features of the implementation :
- 8x8 DCT and IDCT calculation using precomputation of transform coeficients. It does
this in two stages, first for rows and then for columns.
- We have fully implemented Huffman encoding and decoding using table lookup for
run-category values.
- Bit retrieval and insertion using buffers.
Implementation Details :
Results
Original Image |
Compressed Image |
Decompressed Image |
Size 11267 Bytes |
Size 13614 Bytes |
Size 12908 Bytes |
Size 13497 Bytes |
Size 12056 Bytes |
Size 14130 Bytes |
|