A short quiz on intermediate-level vision

Adrian F. Clark

1. What is hysteresis thresholding?

   • Using two thresholds to partition an image into three levels
   • Using hysteresis to choose the value of a threshold
   • Using two thresholds to partition an image into two levels

   • Joining together edge segments

   The last stage of the Canny edge detctor, hysteresis thresholding is a technique for joining together edge segments.

2. What is the profile of the mask used in the first stage of Canny?

   • Sobel
   • Laplacian
   • 3 x 3 blur

   • Gaussian

   The first stage of Canny involves convolving with a mask with a Gaussian profile.

3. What is the underlying principle of the Moravec corner detector?

   • moving along an edge introduces a large change
   • moving away from a corner introduces a small change
   • moving away from a uniform region introduces a large change

   • moving away from a corner introduces a large change

   The Moravec corner detector attempts to identify regions where a movement in any direction increases the response.

4. What type of feature is LBP best at identifying?

   • different types of texture

   • edges
   • faces
   • corners

   LBP, Local Binary Patterns, is normally used to describe texture.

5. What is the purpose of the Hough transform?

   • To find edges in images

   • To find the parameters of lines, etc

   • To join together edges in images
   • To calculate an accumulator array

   The purpose of the Hough transform is to estimate the parameters that describe some image feature -- though we have looked only at straight lines in this module.

6. Which of the following is not really an edge detector?

   • Harris and Stephens
   • Sobel

   • Laplacian

   • Canny

   Sobel and Canny were designed as edge detectors. Although better known as a corner detector, Harris and Stephens also detects edges. On the other hand, the Laplacian is better at detecting isolated points.

7. In the non-parametric line Hough transform, what is the accumulator array indexed by?

   • end-points
   • radius and angle
   • gradient magnitude and direction

   • gradient and intercept

   The non-parametric Hough transform estimates the gradient and intercept directly, so we use these as the indices of the accumulator array.

8. The Canny edge detector represents more or less the state of the art in detecting edges in images. Why is its output still of limited use?

   • it is too slow
   • it doesn't find all edges

   • edges are of limited value in practice

   • the edges are still too noisy to be useful

   In most cases, identifying the edges of objects is useful for segmenting them from the background; but edges are not much help in (say) identifying the object. For that task, it is much better to use the corners of features in the image.

9. What shape is the curve drawn in the parametric accumulator array of a Hough transform from a single pixel in an image?

   • an $x^2$ shape

   • a sinusoid

   • a $1/x$ shape
   • a straight line

   The parametric form of a straight line is a sinusoid, so that is what is drawn into the accumulator.

10. What is the purpose of the lower threshold in hysteresis thresholding?

    • Pixels with values below this threshold are not parts of edges

    • Edge segments with values below this threshold are joined to edge segments known to be correct
    • To identify dark lines in images
    • To identify dark regions in images

    Pixels whose values lie below this threshold are rejected as edges, while pixels whose values lie above this threshold are potentially edges, depending on the values of their neighbbouring pixels.