An efficient FPGA implementation of Anisotropic Diffusion function for medical images
| Affiliation auteurs | Affiliation ok |
| Titre | An efficient FPGA implementation of Anisotropic Diffusion function for medical images |
| Type de publication | Conference Paper |
| Year of Publication | 2018 |
| Auteurs | Fredj AHadj, Malek J, Bourennane E-B |
| Conference Name | 2018 15TH INTERNATIONAL MULTI-CONFERENCE ON SYSTEMS, SIGNALS AND DEVICES (SSD) |
| Publisher | IEEE; Tunisian Assoc Appl Sci & Technologies; ERE Tunisia; Novelti; Thyna Petroleum Servive; Univ Sfax, Ecole Natil Ingn Sfax; Leipzig Univ Appl Sci; Techn Univ Chemnitz; Philadelphia Univ; King Fahd Univ Petr & Minerals; Univ Sidi Mohamed Ben Abdallah; E |
| Conference Location | 345 E 47TH ST, NEW YORK, NY 10017 USA |
| ISBN Number | 978-1-5386-5305-0 |
| Mots-clés | Anisotropic diffusion, edge preserving smoothing, FPGA, hardware acceleration, real-time systems |
| Résumé | Anisotropic diffusion filter is very useful technique in image processing. This filtering method has been shown to be effective in enhancing the visual quality of medical images and as preprocessing prior to advanced image processing. However, using standard software implementation, it is too hard to perform real-time anisotropic diffusion filtering due to its arithmetic complexity. To overcome this issue, hardware design is the ideal choice. This paper gives the algorithm and implementation of diffusion coefficient calculator on FPGA. The function has been successfully implemented and synthesised using the Xilinx ZYNQ-7 ZC702 FPGA. The platform shows that the diffusion coefficient calculator can operate at a maximum frequency of 210 MHz. The experimental results reveal a 143x and 5x performance improvement over the original c++ implementation and the GPU implementation by using NVIDIA Geforce 840M card, respectively. |