Research Article
A Hybrid Vector Quantization Combining a Tree Structure and a Voronoi Diagram
Input: an input vector x, a root node of tree structure root, and a set of Voronoi regions V | Output: an optimal codeword | Greedy search algorithm: | (1) Project x to , and denote the projected vector as | (2) Call = to find a Voronoi region | (3) Call = to find an optimal codeword with a | predefined number of ripples, , and set the minimal distance is as infinite | (4) Return | Function | (5) If then { | (6) If then | (7) Return a Voronoi region | (8) Else | (9) Call , | (10) } | (11) Else { | (12) If then | (13) Return a Voronoi region | (14) Else | (15) Call , | (16) } | Function , , , | (17) If > 1 then { | (18) Find a such that is minimized | (19) Let be the neighbors in group of Voronoi region | (20) Find a such that the distance, , between and the codeword of is minimized | (21) If then { | (22) , , , | (23) Return | (24) } | (25) } | (26) Return |
|