|
|
|
|
[ HOME ] [ personal ] [ articles ] [ computing ] [ projects ] [ resume ] |
|
|
|
D o w n l o a d * JAMI binary (85kb) | source (58kb)
The JAMI program was written jointly by me and Jai Vasanth (a classmate at SVCE). The huffman tree is one of the most well known compression methods which allocates based on frequency of codes. The trick is in implementing easy to use tree structures and efficient compression techniques. The compressor has one problem namely the use of a count rather than probability. However as the count variable is 64-bit, I do not foresee any problems. The compressor also caches 16kb of data when compressing. The bit structure class was a pain to implement and Jai Vasanth deserves credit for the design. The bit structure class dumps a byte everytime 8 bits have been used. This can be used to write the compressed code. The decompressor is slow mainly because I haven't been able to find any method besides traversing the huffman tree in order to decode. One of my friends suggested that I should use an array structure rather than dynamic array for the tree.
|
