A quiz on image representation and processing in software

Adrian F. Clark

1. In OpenCV, if Y indexes lines, X columns and C channels, the right nesting >of loops is

   • X inside Y inside C

   • C inside X inside Y

   • Y inside X inside C
   • C inside Y inside X

   The C subscript should cycle fastest in all languages you will use OpenCV with, followed by the X subscript. The Y subscript should cycle slowest. This order mimics the way in which the pixels are stored in memory and thus minimises cache misses etc.

2. What is plotted along the abscissa (x-axis) of a histogram?

   • number of occurrences

   • grey level

   • standard deviation
   • mean value

   The x-axis of a histogram plots the grey level; the y-axis plots the number of occurrences of that grey level.

3. Why is the order of subscripts into an image important?

   • it affects the run-time of an algorithm

   • it affects the amount of memory used
   • it makes mapping code onto a GPU easier
   • it makes the code easier to read

   When the subscripts are indexed in the right order, adjacent pixels in the image are adjacent in memory. This reduces the number of cache hits, allows auto-increment instructions to be used, and so on.

4. What range of values can be stored in an unsigned byte?

   • -127 -- +127
   • -128 -- +127

   • 0 -- 255

   • 0 -- 256

   An unsigned byte represents 256 distinct values, namely 0, 1,... 255.

5. If using histograms to perform content-based image retrieval, which of the following is suitable for determining similarity?

   • the cross-correlation

   • the difference
   • the square
   • the product

   One can use the sum-squared difference but not the difference, product or square. The cross-correlation, however, does determine similarity.

6. Roughly how much memory does a broadcast-size colour image require in EVE?

   • 2.5 Mbyte
   • 0.5 Mbyte
   • 1.2 Mbyte

   • 5 Mbyte

   A PAL image comprises 576 (visible) lines of 720 pixels, each of which has red, green and blue components (after unpacking: the actual broadcasts use YIQ format, which reduces the amount of colour information transmitted). EVE uses floats, which are 4 bytes in size. So the total size of an image is 720 x 576 x 3 x 4, which comes to about 5 Mbyte.

7. How does histogram equalization work?

   • histogram bins with low occupancy are doubled
   • histogram bins with low occupancy are merged
   • histogram bins with low occupancy are deleted

   • histogram bins with low occupancy are spread out

   Histogram equalization spreads out low-occupancy parts of a histogram. It is inhereently non-linear in nature and generally a very silly thing to do for computer vision. Your lecture can count on the fingers of one foot the number of times it has helped him develop vision algorithms.

8. Which of the following is not a TV broadcast standard?

   • GRL

   • NTSC
   • PAL
   • SECAM

   NTSC is used in the USA and Japan, SECAM in France and Russia, and PAL in most of the rest of the world. GRL stands for Group of Random Letters.

9. Why is imagery with 8 bits per pixel commonly used in vision software?

   • 8-bit values are faster to process than integers
   • 8-bit values fit into a byte

   • the human eye sees less than 8 bits worh of grey levels

   • computers don't have enough memory to use more than 8 bits/pixel

   The underlying reason is that the eye sees less than 256 shades of grey, though they are not evenly spaced in intensity. It is a convenient coincidence that this matches one byte per pixel. Depending on the hardware involved, 8-bit values are processed at the same speed as integers. Unless it is extremely large (and some of those your lecturer processes are), images are small compared to the size of memory these days.

10. OpenCV was initially developed in

    • Australia
    • USA

    • Russia

    • UK

    Perhaps surprisingly to some, the early versions of OpenCV were developed in Russia, at Intel's research center in Nizhny Novgorod.