Abstract
Image compression using JPEG algorithm has revolutionized the digital multimedia industry. JPEG based image compressions requires lower bandwidth for transmission and reduce storage disk space. Bitmap Image Format stores the pixel values without any encoding or compression. Hence it has larger size than the JPEG file format. Decompression of JPEG image involves extracting the pixel values using 2-dimensional Inverse Discrete Cosine Transform and de-quantization. The pixel values are in zigzag format in JPEG so they need to be extracted out in the normal format. JPEG header has discrete quantization tables and Huffman tables encoded in it. These tables need to be extracted. Apart from this, bitmap header format extraction and reorganizing them along with the pixel values is needed for converting a jpeg format image to a bmp format image. This project explores the various possible architectures for hardware software co-design. The project implements Hardware/Software Co-Design using Atmel ATmega32 micro controller as hardware and Visual C++ is used to implement the software part of the project. Hardware software co-design combines the best of both. Software gives the flexibility in design while the hardware guarantees the performance, throughput and efficient operation.