Abstract

The proposed TSIHT coding keeps low bit-rate quality as SPIHT algorithm and has three improved features. Firstly, to reduce the amount of memory usage, TSIHT coding introduces tag flags to store the significant information instead of the coordinate-lists in SPIHT. The tag flags are four two-dimensional binary tag arrays including Tag of Significant Pixels (TSP), Tag of Insignificant Pixels (TIP) and Tag of Significant Trees (TST) respectively. When comparing with SPIHT coding, TSIHT only needs 26 K bytes memory to store four tag-arrays for a 256×256 gray-scale image. Secondly, both sorting-pass and refinement-pass of SPIHT coding are merged in one in TSIHT coding in order to simplify hardwarecontrol and save unnecessary memory.

Finally, TSIHT uses the Depth-First- Search (DFS) traversal order to encode bit-stream rather than the Breadth-First-Search (BFS) method as the SPIHT coding. Since, DFS method searches te root node and each one of the branching node of the immediate descendants until it reaches the deepest leaves. For the hierarchical pyramid nature of the spatial orientation tree, DFS provides a better architecture than BFS method. Additionally, a VLSI image compressor called PIE (Progressive Image Encoder) core for TSIHT coding has been implemented onto a chip with 0.38 μm one-poly-four-metal CMOS technology

LANGUAGE USED:

VHDL

TOOLS REQUIRED:

• Simulation: ModelSim XE III 6.4b.
• Synthesis: XiLinx ISE 10.1.

<< back