Contrast Based Enhancement of Palm-Leaf Manuscript Images

This paper presents a new color document image processing scheme, suitable for enhancement of palm-leaf manuscript images. Generally, image analysis systems consist of three stages: pre-processing, binarization and post-processing. In the proposed scheme, pre-processing is done using a newly developed contrast based RGB to grey-scale conversion technique and other two stages are combined in a novel two-level global binarization technique. First, the input RGB color image is converted into its enhanced grey-scale version on the basis of relationship between contrast ranges of three color planes, R, G and B. The threshold for two-level global binarization technique is decided on the basis of contrast range and mean of gray-scale image. Use of enhanced gray-scale image in proposed and enhanced binarization techniques shows significant improvement in the quality of binarization results.