Effective restoration of clipped pixels in LDR images

Conventional images store a very limited dynamic range of brightness. When conversion of HDR images to LDR images takes place for making images compatible to low dynamic range display, some part of HDR images get clipped. The clipping changes the value of RGB results, color distortion occurs. For getting clipped pixels we propose an algorithm to enhancing the chroma values. Proposed algorithm is mainly based on chroma spatial correlation between clipped pixels and surrounding unclipped areas. After identifying the clipped area in image, partition the clipped area with similar chroma, then estimate the chroma for clipped area by the help of neighboring unclipped chroma pixels. Correct the RGB values by the help of estimated chroma values of the clipped area. Last step is to smooth the boundary where sharp color difference occurs, by providing transition band. When result is compared with the original image, it shows that algorithm is very effective in restoring the clipped pixels.