To Study and Implementation of Image Compression Using Inexact Computing

This paper suggests a new method for digital image processing; it focuses on incorrect computation to solve some of the problems associated with discrete cosine transformation (DCT) compression. The suggested system has three stages of processing; the first stage uses estimated DCT for image compression to remove all computing intensive floating-point multiplication and to perform DCT processing with integer additions and, in some situations, logical right / left changes. The second level also decreases the volume of data (from the first level) that need to be processed by extracting the frequencies that cannot be sensed by human senses. Finally, in order to minimize power consumption and latency, the third stage incorporates incorrect circuit level adders for computing the DCT. For measurement, a collection of structured images is compressed using the proposed three-level system. Different figures of merits (such as energy consumption, delay, power-signal-to-noise-ratio, average-difference, and absolute-maximum-difference) are compared to existing compression methods; an error analysis is also pursued confirming the simulation results. Results show very good improvements in reduction for energy and delay, while maintaining acceptable accuracy levels for image processing applications.