Quantity control for JPEG image data compression using fuzzy logic algorithm

The JPEG established the first international standard for continuous-tone still image for both gray scale and color. However, since the JPEG's standard was originally designed for general applications, some modifications must be done for specific applications such as in digital still cameras. We propose a new method to control the bit-rate of the JPEG standard using a fuzzy logic algorithm (FLA). A gain factor is determined to derive an appropriate quantization matrix. Several rules have been prepared to calculate the gain factor. We used 18 test images for simulations and the results show that the mean error is -1.5%, standard deviation is 1.5% and the error range from -3% to 1%. Without the FLA, the error range from -41 to 43%, the mean error is -0.11% and the standard deviation is 24.24%.<
>