LL6

Question 1

What is segmentation in image analysis?

Answer 1

Segmentation is the process of dividing an image into foreground and background, subdividing it into constituent regions and/or objects, and transforming an image of densities to a binary image.

Question 2

What are the two basic properties of intensity values used in segmentation algorithms?

Answer 2

• Discontinuity
• Similarity

Question 3

What does the term ‘local adaptive enhancement’ refer to?

Answer 3

Local adaptive enhancement considers a window to improve local effects and can be applied with histogram equalization.

Question 4

What is CLAHE and what are its benefits?

Answer 4

Contrast Limited Adaptive Histogram Equalization; it is less sensitive to noise and enhances contrast.

Question 5

What is the purpose of thresholding in image segmentation?

Answer 5

Thresholding selects pixels of intensity range to produce a binary image.

Question 6

What is bi-level thresholding?

Answer 6

Bi-level thresholding involves selecting a lower threshold T1 and an upper threshold T2 to classify pixels.

Question 7

What is the Isodata method in segmentation?

Answer 7

Isodata is a thresholding technique developed by Ridler & Calvard, starting with a threshold value and iteratively computing means of foreground and background.

Question 8

What does entropy represent in image analysis?

Answer 8

Entropy represents information richness; a high number indicates that the dynamic range is fully utilized.

Question 9

Fill in the blank: The _______ method is a local thresholding technique that uses local average and standard deviation.

Answer 9

Niblack

Question 10

What is the primary challenge of segmentation techniques?

Answer 10

There is no perfect segmentation technique; a technique must fit the data.

Question 11

What is the significance of the histogram in image analysis?

Answer 11

The histogram represents frequency distributions of intensities and can be used to enhance image contrast.

Question 12

What does the term ‘grouping’ refer to in the context of segmentation?

Answer 12

Grouping is the process of associating similar image features together.

Question 13

What is meant by ‘local contrast enhancement’?

Answer 13

Local contrast enhancement improves contrast in specific regions of an image using localized methods.

Question 14

True or False: Segmentation can be used to obtain a mask for image analysis.

Answer 14

True

Question 15

What are the limitations of thresholding?

Answer 15

Illumination is often not uniform, affecting the applicability of thresholding, which may require adaptive methods.

Question 16

What is the purpose of illumination correction in microscopy?

Answer 16

Illumination correction addresses shading due to uneven illumination from lamp imperfections.

Question 17

What is the role of the ‘dark current’ image in correction processes?

Answer 17

The dark current image is used to correct for sensor noise in image analysis.

Question 18

What does the term ‘density slicing’ refer to?

Answer 18

Density slicing is a technique that selects pixel intensity ranges to create binary images.

Question 19

What is ‘contrast stretching’?

Answer 19

Contrast stretching is a linear method to enhance image contrast.

Question 20

What is the Gamma function used for in image processing?

Answer 20

The Gamma function is a non-linear method used to enhance image contrast.

Question 21

What is the ‘Otsu’ method?

Answer 21

Otsu’s method is a thresholding technique that determines the optimal threshold to separate foreground and background.

Question 22

What is meant by ‘region growing’ in segmentation?

Answer 22

Region growing is a method that selects similar pixels to form larger regions based on predefined criteria.

Question 23

What is the purpose of annotation in image analysis?

Answer 23

Annotation involves marking specific features or regions of interest within an image for further analysis.

Question 24

What is Niblack’s method used for?

Answer 24

Local thresholding in image analysis

Niblack’s method was introduced in 1986 for local adaptive thresholding.

Question 25

What parameters are involved in defining the local environment in Niblack's algorithm?

Answer 25

Local average and local standard deviation ## Footnote The local environment is typically defined by a window size, denoted as r*r.

Question 26

What is the formula for the threshold T(x,y) in Niblack's algorithm?

Answer 26

T(x,y) = .(x,y) + k * .(x,y) ## Footnote This formula incorporates the local average and standard deviation at the pixel (x,y).

Question 27

In Niblack's algorithm, what does the parameter 'k' typically equal?

Answer 27

-0.2 ## Footnote This value is used to adjust the thresholding based on local variations.

Question 28

What is Bernsen's method known for?

Answer 28

Local adaptive thresholding ## Footnote Introduced in 1986, this method computes thresholds based on local contrast.

Question 29

What does the contrast C(x,y) represent in Bernsen's algorithm?

Answer 29

C(x,y) = Zhigh - Zlow ## Footnote Zhigh and Zlow are the highest and lowest pixel values in the local window E.

Question 30

What condition must be met to classify a pixel as background in Bernsen's algorithm?

Answer 30

C(x,y) < l ## Footnote Where 'l' is a threshold value that depends on image quality.

Question 31

What is the primary goal of Otsu's method?

Answer 31

To find a threshold value that minimizes the sum of foreground and background spread ## Footnote Otsu's method involves iterating through all possible threshold values.

Question 32

What does the Within-Class Variance represent in Otsu's method?

Answer 32

The variance of pixel levels within each class (foreground and background) ## Footnote It is computed to help find the optimal threshold.

Question 33

What is the significance of the parameter 'r' in local adaptive methods?

Answer 33

It defines the size of the local window or kernel ## Footnote Both Niblack and Bernsen methods use a kernel of size r*r.

Question 34

What does region growing in image analysis require?

Answer 34

A seed pixel and a similarity criterion ## Footnote The seed pixel initiates the growth based on its density value and location.

Question 35

What is the stopping criterion for region growing?

Answer 35

If neighbors are outside a given range ## Footnote The range is defined by Tlow and Thigh, derived from the seed pixel.

Question 36

What does K-means clustering do?

Answer 36

Partitions an image into k clusters ## Footnote Each cluster is represented by the mean of its data points.

Question 37

What is a common issue with K-means clustering in image segmentation?

Answer 37

It may not completely resolve segmentation ## Footnote Additional heuristics may be needed to address unconnected parts.

Question 38

What is the purpose of annotation in image segmentation?

Answer 38

To attribute labels to each segmented part ## Footnote Labels can describe the context of the visualized object.

Question 39

What are heuristics in the context of segmentation?

Answer 39

Rules that depend on the image to guide segmentation ## Footnote They help in understanding the syntax of the scene.

Question 40

What is the Hough transformation used for?

Answer 40

Line and circle detection in images ## Footnote It is a technique often used in edge detection.

Question 41

What does the term 'superpixels' refer to in image analysis?

Answer 41

Regions that are perceptually meaningful ## Footnote They represent groups of pixels rather than individual pixels.