Physics, Psychophysics and Physiology of Vision Practical Practical One
- Colour Imaging
10%, to be completed by 7/2/2003
1. Objective
In the lectures, you have heard about different colour models or
representations and about 'colour constancy'. The aim of this practical
is to illustrate some of these concepts and to exercise some of the
techniques used in the colour image processing.
2. Procedure
Carry out each of the steps listed below and, as in the Machine Vision
practicals last term, write a brief report containing:
-
a description of the techniques used, where appropriate in succinct
mathematical terms
-
a description of what was done, what data was used, how it was obtained,
what tools, library facilities etc were used
-
the results obtained
-
an analysis and critique of the results
-
any conclusions you can draw from the exercise
Include any code you write yourself (suitably commented) in an appendix.
3. Exercises
Three sets of images, all in ppm format, may be copied as required from
the directories:
-
/cs/research/vision/images/daedalus/Cloth1 (mustard coloured
shirt, imaged under various lighting conditions),
-
/cs/research/vision/images/daedalus/MCloth5 (images of several items of
clothing, including a patterned jumper, a sock and a towel), and
-
/cs/research/vision/images/daedalus/Track2 (images of a country lane in the
South Downs).
Select five examples from each set for the exercises below. Note that
many of the images are rather dark, so it is best to view (but not
process) the originals after histogram equalisation, for example using
xv.
3.1 Image Data
Use xv, any other convenient tools, or write your own procedure in IDL to crop
or define regions of interest on:
-
the mustard coloured shirt in five images chosen from the first clothing
sequence,
-
the sock in five images chosen from the second clothing
sequence,
-
the country lane in five images chosen from the outdoor sequence.
It is important to ensure, especially if a cropping tool is used to
define the regions by hand, that the selected regions within each
image set are exactly the same size, P x Q. It is thus best to use the
crop tool and mouse in xv to define the regions of interest and then
to select them using IDL. If necessary rows or columns may be deleted
from the requisite images afterwards. If you find it difficult to
select cropped regions from the images themselves, the white templates
(NB. not the grey which is a region of uncertainty that may or may
not be part of the object) in the images in:
- ../Cloth1GT,
- ../MCloth5GTs,
- ../Track2GT,
define the requisite objects in each correpondingly named
image.Variations in size of the selected regions between sets are
unimportant and it does not matter if the selected regions correspond
to different parts of the same object, athough they should be
representative of the object.
Use xv and IDL to convert image formats and to extract the RGB
components of the images so they can be accessed as matrices for
processing as described in sections 3.2-3.5 below. See the notes
prepared earlier by Ioannis Douros for the second MV practical on image
preparation for further details, in particular on how to set up image
matrices for each colour channel. As usual, it is probably best to
perform all of the operations indicated on one of the images or image
sets first in order to be sure that your implementation works and to
minimise the number of images and intermediate results you have to
store.
3.2 Colour variation in RGB space
Compute the mean colour of each chosen image in RGB space, construct the
scatter matrix and find the principal components of the colour distribution
for each image. Repeat the calculation of the mean and principal components
using data from all the chosen images within each of the three sets (ie
take the mean over all five of your chosen images etc.).
View the images and image regions used and consider the variations in
the results of means and PCA within each set, comment on them in light of
what you would expect from Schafer's dichromatic reflectivity model and
indicate what the results tell you (if anything) about the lighting
conditions under which each of the three sets of images was taken.
If time allows, it may help to plot the mean colour and a
representation of the principal components of variation; eg the
principal (eigen)vectors, the eigenvectors scaled by the standard
deviation along that vector, or an ellipsoidal surface of the
corresponding Gaussian probability distrubtion. Remember, if you do
this, to use the mean colour as a reference yfor plotting the
components of variation or a surface of the probability distribution.
3.3 Colour (row) normalisation and analysis of chromaticity
Compute the normalised colour representastion rgb of each of the chosen
images, calculate the mean rgb colours or chromaticity and repeat the PCA
analysis in the rgb space. If the images are regarded as PQ x 3 rectangular
matrices as in the work of Finlayson et al (1998), colour normalisation
corresponds to row normalistion of the rectangular matrices.
One of the principal components should now be zero with its axis oriented
along the white direction. Explain why this should be so and check that
your results are consistent with it. View the images resulting from the
normalisation, compare the results of the PCA obtained in the rgb space with
those obtained previously and comment on what you find. Be as quantitative
as possible in making this comparison.
If time allows, investigate what happens if a Euclidean norm is used
to define the normalised colours or chromaticities.
3.4 Pixel (column) normalisation
Normalise each original RGB image within each colour channel by summing
over all the pixel values and, as in the work of Finlayson et al (1998),
scaling the result by a factor of PQ/3. Calculate the mean colours and
principal axes of these normalised images as in 3.2 above, view the resulting
images and comment on the results obtained.
Similarly, normalise and scale the images in rgb representaion as obtained
in section 3.3 above. Use only the conventional normalisation R+G+B for
this step. VIew the images resulting from this normalisation and comment
on the results obtained both in comparison to those from the unmodified RGB
images imediately above (this section, 3.4) and in comparison to those from
the rgb (row) normalised images in 3.3 above.
3.5 Comprehensive normalisation
Carry out a comprehensive normalisation of the images as described in the
work of Finlayson et al (1998) by iteratively repeating the row normalisation
described in section 3.3 and the column normalisation in section 3.4.
Comment on the convergence of this iterative procedure and view the
comprehensively normalised images. Calculate the mean colour of the
comprehensively normalised images and carry out a similar PCA analysis of
their distributions in colour space to those described above. Comment on
the results obtained and on the potential significance of each of the
three types of normalisation procedure.
4. Reporting
As usual, beware of labouring long and hard to produce a pretty report.
There is a rapidly diminishing rate of return for such efforts. Evidence
that the exercise has been completed and content of the report is much
more important than presentation. Adequate presentation and analysis of
the original and normalised images should however, be included in your
report, as part of this evidence.
Bernard Buxton,
14 January 2002.