Efficient progressive JPEG decoder using JPEG baseline hardware
Mihir Mody, Vipul Paladiya, Kapil Ahuja · 2013
The Widespread usage of social media for picture sharing led resulting popularity of JPEG progressive format due to refinement of image over time on slow internet connection. Typically, these pictures are decoded by means of software and takes large decoding time as resolution in terms of Mpixels increase. In case of embedded devices, typically have Baseline JPEG hardware support due to traditional camera capture and local playback. The paper proposes a novel decoding of Progressive JPEG images using existing hardware solution which natively supports Baseline JPEG. The solution used enhancing native baseline JPEG hardware by means of local CPU to enable Huffman decoding for Progressive format. The second part of part proposed memory optimization for Progressive JPEG to reduce DDR storage requirement for low cost embedded system. This uses concept of sign and zero map with non-bit exact decoding without visual quality impact. The solution runs faster by large factor (factor of HW and CPU speed) with 88% lesser memory storage for any resolution.