VLSI Design of 2-D Discrete Wavelet Transform for Area-Efficient and High-Speed Image Computing - Mid Presentation Presentor: Eyal Vakrat Instructor: Tsachi Martsiano
Table of content Project goals Project milestones Block diagram Micro-architecture Top level architecture Results so far Gantt Development environments
Project goals Implementation of high-speed and real-time 2-D Discrete Wavelet Transform Based on new and fast convolution approach Efficient memory area (in-place) Article I use: World Academy of Science, Engineering and Technology 21 2008, VLSI Design of 2-D Discrete Wavelet Transform for Area-Efficient and High-Speed Image Computing, by Mountassar Maamoun, Mehdi Neggazi, Abdelhamid Meraghni, and Daoud Berkani.
Project milestones Learn the 2D-DWT algorithm from the article Write floating point MATLAB DWT and IDWT Choose coefficients Compare the results to MATLAB DWT function Write fixed point MATLAB DWT and IDWT Select the fixed point resolution Architecture: Learn the proposed architecture from the paper Adjust it to our case - different coefficients and picture size Code the module in VHDL Simulate the module using ModelSim Synthesis of the module using Vivado
Top Block Diagram memory
Micro-architecture
Top Level Architecture
Results so far Performing the DWT in Matlab In progress: Writing the DWT pipe line and testing it Choosing the Memories
Results so far – cont. Original picture: My result(float): MATLAB result:
Results so far – cont. Fixed point(1024) Floating point
Gantt We are here!
Development environments MATLAB - modeling MODELSIM -simulation VIVADO - synthesis
THANK YOU!