Implementation of image compression on hardware platform

• Main Author: Koh, Aloysius Jia Lun
• Other Authors: Ho Duan Juat
• Format: Final Year Project
• Language: English
• Published: 2018
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: http://hdl.handle.net/10356/74553
• Institution: Nanyang Technological University

The image compression in which I use is JPEG. JPEG has been around for many decades and it has also been a standardized means of image transmission. For JPEG, there are lossless and lossy compression. Lossy compression will be implemented for this project. Lossy compression will result in a lower quality image as compared to the original, but it requires fewer resources and lower latency. The values and test image are according to industry’s standard and will be implemented as a means of testing the algorithm effectiveness on Zedboard and on a hardware platform. Given that the coding and methods used were simple and not effective, it would only be a rough comparison. “Lena”, a standardized test image was converted to RGB format and adjusted to 256 x 256 resolution, prior to the compression, to accommodate for the low storage space. The image will be converted to YCrCb color space for the JPEG compression and back to RGB, to be displayed through a 12-bit VGA cable to a monitor. The quantization stage of the JPEG compression algorithm determines the quality of the decompressed output image. The higher the quality factor, the lower the compression but higher the output quality. Multiple quality factors in a factor of 10 from 0 to 100 are being tested. However, there are some setbacks. The internal memory of the BRAM is not large enough to support a Q-factor of above 80. As a result, only Q-factor from 0-80 are manipulated.