Estimation of Minimum Quantization Levels by Using Reconstructed Histogram

The OK-quantization theory determines the minimum gray level by using the reproducibility of an image histogram. In many cases, it is ascertained by the human sense of sight that the minimum gray level obtained from this theory is appropriate. However, in order to put the OK-quantization theory into practical use, it is necessary to perform a validity evaluation of this theorem using a computer algorithm. In this research, the gray level of each pixel of a quantized image is first interpolated using the sampling function to create a reconstructed image with the average gray level of each pixel being as a new gray level. The gray level is then evaluated for validity by comparing the histogram of this reconstructed image with that of the original image. The experiment results have confirmed that the proposed method makes it possible to replace an estimation of minimum quantization levels that relies on the human sense of sight, with a computational algorithm.