A quiz on the histograms laboratory

Adrian F. Clark

1. The initial histogram routine you developed plotted values in what range?

   • between the values returned by calls to lowest and highest

   • zero to the value returned by a call to highest

   • zero to 255
   • between the values returned by a call to extremes

   The initial histogram routine fixes the lower limit of the plot at zero and goes up to the value returned by calling highest.

2. The routines you are writing for sxcv.py are specified by their documentation and some tests. What form of extreme programming is this?

   • pairwise programming

   • test-driven development

   • web-search oriented
   • trial and error

   It is common in industry for people to write tests for their code before embarking on writing it. Doing so generally produces correct code more quickly than writing the tests afterwards (because people tend to test the code they have written rather than what the code is trying to achieve) and VASTLY faster than not writing tests at all!

3. How many distinct peaks did the histogram of fruit.png show?

   • 1
   • 3
   • 0

   • 2

   There are two peaks, a large one due to the background and a smaller, broad one ffrom the actual fruit.

4. Which of the following best describes the exposure of image 2006-10-23-002.jpg?

   • under-exposed

   • over-exposed
   • well-exposed
   • clearly-exposed

   This image is much darker than it should be, so it is under-exposed.

5. Image 2006-10-23-001.jpg is well exposed. How does its histogram appear?

   • there is a peak at the left of the histogram
   • there is a peak at the right of the histogram

   • there is a peak around the middle of the histogram

   • the histogram shows no peaks

   Well-exposed images tend to have a preponderance of pixel values around mid grey levels, so there is usually a peak in the middle of the histogram. That is certainly the case for this image.

6. Why does your histogram routine need different cases for monochrome and colour images?

   • so that plot_histogram will work correctly

   • colour images require 3 subscripts and monochrome only 2

   • there is no need to have different cases
   • there are three output arrays for a colour image

   OpenCV stores monochrome images as 2D arrays, indexed by row and column. Colour images also have these two subscripts but a third one is needed to select the colour channel too.

7. For a colour image, the histogram is plotted in what colours by sxcv.py routine plot_histogram?

   • only grey
   • cyan, magenta, yellow
   • different shades of grey

   • red, green, blue

   plot_histogram is set up with default colours of red, green and blue for displaying the histogram of a colour image; it uses grey for a one-channel histogram.

8. What are plotted along the x- and y-axes of a histogram?

   • frequency and colour
   • colour and frequency
   • frequency and grey level

   • grey level and frequency

   A histogram is a plot of the number of pixels which have each possible grey level. Conventionallym the grey level is plotted along the x-axis and the frequency count along the y-axis.

9. What is likely to happen if you cann sxcv.binarize with its below and above parameters set to 255 and zero respectively?

   • the image will come out with background pixels set to zero and foreground ones set to 255
   • the image will come out with all pixels set to zero
   • the image will come out with background pixels set to 255 and foreground ones set to zero

   • the image will come out with all pixels set to 255

   There is a problem with the way that most people implement this routine, including your friendsly neighbourhood lecturer at first. If you call cv2.threshold with its above set to zero, all the pixels above the threshold are set to zero and so are the background pixels. At least in my code, all pixels having the value zero are then set to 255, meaning that the whole image becomes 255. Grr.

10. What is the first value returned by cv2.threshold?

    • the grey level used for pixels below the threshold
    • the grey level used for pixels above the threshold
    • the threshoolded image

    • the value of the threshold

    The first value returned is the threshold. It returns this because when you tell cv2.threshold to use use Otsu's algorithm to determine the threshold value itself, you wouldn't otherwise know what it is set to.