Image Lookup Table Compression Using Scaled Partitioning with Memory Data Compression

Methodologies to reduce LUT ROM amount are proposed using input pixel partitioning. To minimize the amount of error in the LUT, a scale factor is used to correct the LUT output values for the partitioned inputs, based on the input values. The dynamic range of lower bit group computation is optimized according to the step size of the upper partitioned bit group computation by the scale factor. A LUT computation and hardware design methodologies are presented utilizing image memory compression. 23.4% of the LUT ROM count is reduced for Kodak image dataset without image quality degradation in histogram equalization for verifications using the presented LUT design methods. According to the image memory compression rate, LUT computation and memory access rate can be significantly reduced.