Description
Description
The goal of this assignment is to gain an understanding of demosaicing: converting the Bayer pixel
pattern into a RGB representation where each pixel has red, green and blue color channels.
Demosaicing is typically the first step in a sequence of non-linear operations of the imaging pipeline that
converts the raw output of the digital camera’s sensor into a JPEG-compressed image which can then be
displayed on the monitor. The figure below shows an example of a Bayer arrangement on the pixel array
of an image sensor.
Figure 1: Bayer pattern
Part 1:
Implement a very simple linear interpolation [averaging the 4 or 2 nearest neighbours] approach that
converts the Bayer arrangement to RGB components for each pixel as shown in the figure below [note that
the figure is an example and does not reflect the Bayer arrangement used by the digital camera]. You
should create a kernel for each channel and use cv::filter2D(…) instead of using loops. The “mosaic” image
was created by taking the original color image and keeping only one color component for each pixel,
according to the standard Bayer pattern:
B R B R. . .
R G R G. . .
B R B R. . .
. . ,
C.POULLIS COMP425-6341 COMPUTER VISION
This method produces artifacts. To visualize these artifacts compute the image of the squared differences
between the original and reconstructed values for each pixel, summed over the three color channels.
Original, demosaic, root squared difference.
Finally, show a close-up of some patch of the reconstructed image where the artifacts are particularly
apparent and explain the cause of these artifacts.
Part 2:
Bill Freeman proposed an improvement of the simple bilinear interpolation approach. Since the R
channel is sampled at a higher rate than the G and B channels, one might expect interpolation to work
better for R values. Then it would make sense to use the interpolated R channel to modify the interpolated
G and B channels. The improved algorithm begins with linear interpolation applied separately to each
channel, just as you have already done above. The estimated R channel is not changed, but G and B
C.POULLIS COMP425-6341 COMPUTER VISION
channels are modified as follows. First, compute the difference images G-R and B-R between the
respective interpolated channels. Mosaicing artifacts tend to show up as small “splotches” in these
images. To eliminate the “splotches”, apply median filtering to the G-R and B-R images. Finally, create the
modified G and B channels by adding the R channel to the respective difference images.
Implement the above algorithm and visualize the quality of the results in the same way as for Part 1 by
displaying the per-pixel difference image. Compare the output to that of Part 1. Are there visible
improvements (especially in the close-up patch selected in Part 1)?
Hint: Implementing this method should take you six lines of code.
Submission (electronic submission through EAS only)
Please create a zip file containing your C/C++ or Python code and a readme text file (.txt).
In the readme file document the features and functionality you have implemented, and anything else you
want the grader to know i.e. control keys, keyboard/mouse shortcuts, etc.
Additional Information
● Mosaic images and their originals can be downloaded from here: image_set.zip
