SKEM4133 MVS 2324-1 Topic 5

Object _________ are the most important cues that link an intensity image with its interpretation.

• shapes

• boundaries

• colors

• textures

An edge detection technique based on the zero-crossings of the _________ explores the fact that a step edge corresponds to __________in the image function.

• second derivative, an edge blurring

• first derivative, an abrupt change

• second derivative, an abrupt change

• first derivative, an edge detection

The ________ of the image function should have an extremum at the position corresponding to the edge in the image, and so the _________ should be zero at the same point.

• second derivative, first derivative

• first derivative, second derivative

• second derivative, third derivative

• third derivative, second derivative

After image convolution with ___________, the locations in the convolved image where the zero level is crossed correspond to the positions of edges.

• derivative of the Gaussian filter

• derivative of the Laplacian filter

• derivative of the Gaussian operator

• derivative of the Laplacian filter

The following statements are TRUE about the Canny edge detector EXCEPT:

• Canny edge detection operator uses a multi-stage algorithm

• It was developed by John F. Canny

• It detects a wide range of edges in images

• It can be used to smooth out sharp edges in images

____________ is an important part of image processing as it facilitates the computational identification and _________ of images into distinct parts.

• Edge detection, separation

• Edge smoothing, separation

• Edge detection, grouping

• Edge blurring, grouping

Choose the correct sequence for the Canny edge detection process:

• Gaussian filtering, Image gradient, Non-maximum suppression, Hysteresis

• Image gradient, Gaussian filtering, Non-maximum suppression, Hysteresis

• Gaussian filtering, Hysteresis, Non-maximum suppression, Image gradient

• Gaussian filtering, Non-maximum suppression, Image gradient, Hysteresis

As noise can mislead the result in finding edges, we have to _____ the noise. Therefore, the image is smoothed by applying a _______________.

• reduce, Laplacian filter

• reduce, Gaussian filter

• eliminate, Gaussian filter

• eliminate, Laplacian filter

Non-maximum suppression is necessary to convert the _______ in the image of the gradient magnitudes to _________.

• blurred edges, sharp edges

• sharp edges, blurred edges

• missing edges, clear edges

• missing objects, sharp edges

In image processing, ________compares two images to build an intermediate image. The function takes two binary images that have been ________ at different levels.

• hysteresis, thresholded

• convolution, smoothed

• threshold, smoothed

• convolution, thresholded