Saturated-pixel enhancement for color images

摘要

We propose an algorithm to correct both luma and chroma of the saturated pixels in an overexposed image. Our method is based on the strong chroma spatial correlation between saturated pixels and their surrounding unsaturated area. We first identify the saturated areas in the image. Then, we partition these areas into regions with similar chroma, and estimate the chroma of each saturated region based on the chroma of its surrounding unsaturated region. Next, we correct the saturated R, G, or B color channels according to the estimated chroma and the unsaturated color channel(s) of the pixel. The last step involves smoothing of the boundaries between regions of different saturation scenarios. Both objective and subjective experimental results show that our algorithm is very effective in restoring the color of saturated pixels.