A quiz on the broken biscuits experiment

Adrian F. Clark

1. What is a sensible threshold on the circularity for a biscuit to be identified as circular?

   • about 15

   • about 12.5
   • $4\pi$
   • about 1

   The smallest possible value is $4\pi$ but the segmenter rarely produces a perfectly circular contour. Experience suggests that a limit of about 15 works on this dataset.

2. How many images contain broken or unusable biscuits?

   • 22

   • 100
   • 6
   • 219

   There are 22 images in the task file which are identified as "rej".

3. How are biscuits distinguished from the background in bikky.py?

   • by colour in HSV space

   • by grey-scale

   • by shape
   • by colour in RGB space

   The program segments foreground from background by thresholding. It might work better if we were to identify the background by colour and then invert the mask...

4. What is a sensible threshold on the rectangularity for a biscuit to be identified as rectangular?

   • about 1.5

   • about $4\pi$
   • 0
   • 1

   Theory suggests that a threshold of about 1 should work but in practice it needs to be larger. About 1.5 works reasonably well.

5. How is a biscuit determined to be broken or unusable?

   • it has a large rectangularity
   • it has a large circularity
   • it cannot be segmented

   • it is not circular or rectangular

   There is no single way to do this but pretty much everyone identifies biscuits that are unbroken, and hence circular or rectangular. Other ones are then broken or unusable.

6. What is the purpose of this experiment?

   • to identify which biscuits are broken
   • to use ways of describing shape
   • to see how to segment images by colour

   • to see how a complete vision system works on real-world data

   Although all the answers are true to some extent, the REAL reason for this experiment is for you to learn that techniques which look as though they should work (like using circularity to identify circular regions) are normally much less neat than the theory would have you believe.

7. What type of problem is this defined as in the task file?

   • label

   • segmentation
   • classification
   • biscuit

   The task file defines the problem as a 'label' one, meaning there is a single label to be determined for each image.

8. How many images are used for training in the task file?

   • 6

   • 100
   • 219
   • 213

   There are only 6 files for training specified in the task file.

9. How do you make bikky.py iterate over the test set in biscuits.task?

   • use -test on the command line
   • edit the task file

   • use -dataset test on the command-line

   • edit the program

   The ability to work on different datasets in the task file is built into bikky.py: you use the -dataset qualifier with the name of the dataset to process.

10. What contour measure must be found for a rectangular biscuit?

    • convex hull
    • axis-aligned bounding box

    • oriented bounding box

    • minimum enclosing circle

    It makes sense to use the bounding box which is aligned with the edges of the biscuit rather than the edges of the image.