Lecture 4 - LZW Compression

Question 1

What based method is this?

Answer 1

Dictionary based

Question 2

What is a dictionary

Answer 2

A collection of strings, with each string having a dedicated codeword to represent it. The codeword is a bit pattern, that can be represented by an unsigned int.

Question 3

What is written to the compressed file?

Answer 3

The bit pattern (codeword). The bit pattern can differ for different chars, as the length changes during runtime of the algorithm.

Question 4

Is dictionary statically available?

Answer 4

No it is built dynamically during both compression and decompression.

Question 5

What does the dictionary intially contain?

Answer 5

All possible strings of length 1.

Question 6

What is ‘closed under prefix’?

Answer 6

all prefixes of current string s are reprsented

Question 7

What is the optimal representation of the dictionary?

Answer 7

A trie due to the closed under prefix property.

Question 8

How many more strings can we store in the dicionary if the available codeword length increases by 1.

Answer 8

The size doubles

Question 9

The 3 LZW Variants:

Answer 9

Constant codeword length: fix the codeword length for all time
-> the dictionary has fixed capacity: when full, just stop adding to it

Dynamic codeword length (the version we have been taught)

LRU version: when dictionary full and codeword length maximal
-> current string replaces Least Recently Used string in dictionary

Question 10

Does the decompression stage build the same dictionary as the compression stage?

Answer 10

Yes, but one step out of phase.

Question 11

Complexity of compression and decompression?

Answer 11

O(n) , as we pass through the text once and populate dictionary at every iteration