Image Compression and Decompression with Configurable Architecture on FPGA platform

Images forms a dominant percentage of information available on the internet and on the storage devices. Compression of images saves bandwidth in transmission and saves storage space and bandwidth. A hybrid method is proposed to compress images on Field Programmable Gate Array using Haar Transformation followed by Autoencoder. The Haar Transformation is implemented on FPGA to attain higher throughput compared existing architecture. In addition, Auto-encoder reduces the distortion and provides a higher throughput while consuming minimum power. The result is compared to JPEG 2000.