Color space transformation (or color correction) needs to be
performed in typical imaging devices because the spectral
sensitivity functions of the sensors deviate from the desired target
color space. Several researchers have shown that when the color
channels are correlated, color correction can result in annoying
sensor noise amplification. We developed a color correction method
that significantly alleviates this problem of noise amplification.
The key idea is to use spatially varying color correction that
adapts to local image statistics.
Summary of the method (Patent Pending)
There is a trade-off between color fidelity and the amplification
of noise. By loosening the constraint of having a fixed color
correction matrix for the entire image, a better trade-off can be
obtained. The summary of the baseline procedure is given below.
- Divide the image into non-overlapping blocks (e.g. 8 by 8).
- For each block, compute the correlation matrix of the R, G and
B channels and estimate the correlation matrix of the image
- Compute the color correction matrix using the correlation
- Apply the newly calculated color correction matrix to all the
pixels in the block.
- Proceed onto the next block and repeat previous steps for that
For more technical details, please refer to the following paper:
Suk Hwan Lim and Amnon Silverstein, “Spatially Varying Color
Correction Matrices for Reduced Noise, ”accepted to IS&T
Color Imaging Conference 2004, Scottsdale, AZ, November 2004.
Color correction examples
Check out some color corrected
images with the new method compared with those corrected with
For more information, please contact Suk Hwan Lim (email@example.com)
at Imaging Technology Dept., HP Labs.