Compressed Sensing Based Fixed-Point DCT Image Encoding

Qixin Wang

Abstract


Compressed Sensing, a booming topic in the field of signal processing, requires little computation when encoding. And fixed-point encoder always performs faster and enjoys much lower device cost. In this paper, a Compressed Sensing based fixed-point DCT image encoding system is proposed to fulfill the device cost and computation requirements of distributed sensor network. In the system, the quantization process, a step required by many DCT based encoding algorithm such as JPEG, is eliminated in that quantization will not improve compression rate in Compressed Sensing based system and is testified to reduce the performance of CS based encoder. Also, when using fixed-point, we gain a balance between representation precision and range by multiplying or dividing the original number with well-choose constants, which are chosen according to mathematical analysis. Experimental results verify the performance of our system. The fixed-point system perform as good as float number representation one and elimination of quantization step also improve system performance much.

Keywords


Compressed Sensing, DCT, Fixed-Point Encoder

Full Text:

PDF

Refbacks

  • There are currently no refbacks.


Creative Commons License
This work is licensed under a Creative Commons Attribution 3.0 License.

ACMA©: World Science Publisher United States