[P1] Structuring of Contourlet Transform for Pipeline-Based Implementation
[P1] Structuring of Contourlet Transform for Pipeline-Based Implementation
Project Description:
The Contourlet transform (CT) is a state-of-the-art, multi-resolution image representation technique that effectively extracts critical information from images. Comprising a low-pass filter (LP) and directional filter bank (DFB), CT finds wide-ranging applications in image and video processing, such as quality assessment, watermarking, denoising, and compression. As CT's demand for real-time signal and image processing applications increases, there is a growing need for its hardware implementation. As far as we know, this is the first comprehensive hardware architecture for CT.
In this project, we propose a novel CT architecture, comprising LP and DFB, that is ideal for efficient hardware implementation. We successfully achieved real-time performance, operating at 76MHz, equivalent to 40 frames per second, and validated our design through FPGA implementation.
Hardware Implementation:
Described in VHDL and then implemented on Altera Stratix II EP2S180F1020CS FPGA device.
Please refer to the following publication for further details:
[J1] S. M. A. Zeinolabedin, N. Karimi, S. Samavi, and T. Kim, "Structuring of Contourlet Transform for Pipeline-based Implementation," Circuits, Systems & Signal Processing (CSSP), pp. 1-24, 2015. (link)