Definitions Flashcards
What is the Euler number in the context of binary image analysis?
The Euler number describes the number of connected components and holes in a binary image. This concept is used in image processing to identify and classify the objects in a binary image.
Explain, in detail, Contrast Stretching. Outline a specific example of how the operation is used in practice
- Improves the contrast of an image by expanding the intensity range of the pixels in the image.
- Commonly used to enhance the visibility of details in an image.
- e.g. Could be used to improve the visibility of small objects in an ultrasound image.
1. first, the intensity range of the pixels would be calculated.
2. A new intensity range would be defined that covers a wider range of values, such as 0-255 for an 8-bit image.
3. The intensity values of the pixels in the image would be mapped to the new intensity range, resulting in an image with improved contrast and better visibility.
Explain, in detail, Image Decovolution. Outline a specific example of how the operation is used in practice
- Restores the original sharpness of an image that has been degraded by blur or noise.
- Commonly used to improve the quality of an image.
e.g. Used to remove blur from an image of a person’s face.
1. A model of the blur that has degraded the image would be created (gaussian blur).
2. An algorithm would be used to invert the blur model and apply it to the degraded image.
This would remove the blur from the image and restore its original sharpness.
Explain, in detail, Histogram Matching. Outline a specific example of how the operation is used in practice
- Used to match the intensity distributions of two images.
- useful for comparing or combining them into a single image.
e.g. Align the intensity values of 2 images taken of the same scene at two different times of the day.
1. Intensity histograms of the two images would be calculated and compared.
2. Intensity values of one of the images would be adjusted to match that of the other image.
Roberts edge detection
Roberts approach: Involves convolving the image with 2 kernels, one for horizontal edges and one for vertical. The kernel has values 1,0,0,-1 and 0,1,-1,0. The convolution is performed by sliding the kernel over the image and calculating the dot product between both. The resulting values are then thresholded to identify the edges in the image.
c(I,j)=sqrt(R1^2 +R2^2)
Prewitt Edge detection
Prewitt approach: Similar to Roberts, but uses different kernels. -1,-1,-1,0,0,0,1,1,1 and -1,0,1,-1,0,1,-1,0,1. The convolution is performed in the same way as the Roberts approach.
c(I,j)=sqrt(P1^2 +P2^2)
Explain region labelling by area
a method of labelling the regions in a binary image based on the area or size of the regions. Larger regions are given higher labels and smaller regions are given lower labels. Useful for identifying and classifying objects in an image based on their size, as larger objects tend to be more important or significant than smaller objects.
Explain region labelling by location
A method of labelling the regions in a binary image based on the location or position of the region. Regions closer to a reference point / region are given a higher label and regions that are farther away being given lower labels.
Explain, in detail, Reconstruction by Dilation. Give an example of how this operation is used in image processing / analysis applications.
Mathematical morphology operation that is used to restore or reconstruct the original shape of an object in an image. Performed by first applying dilation to the object to expand its shape, and then applying a mask to the dilated image that keeps only the pixels that are inside the original shape of the object.
Can be useful for removing noise or small, unwanted objects from an image while preserving the original shape of the object.
e.g. removing small, isolated foreground pixels from a binary image of a fingerprint.
Explain, in detail, Greyscale Dilation. Give an example of how this operation is used in image processing / analysis applications.
The maximum for which the structuring element at that location does not fit entirely within the background above the surface. Used to expand the shape of an object in a greyscale image.
Explain, in detail, Geodesic Dilation. Give an example of how this operation is used in image processing / analysis applications.
Point-wise minimum between a mask image and the elementary dilation of the marker image.
Used to expand the shape of an object in an image based on the predefined metric or distance function. Similar to regular dilation, but it allows for more control over the shape of the dilated object by using a distance function to define how the pixels in the image should be expanded.
e.g. Used to expand the shape of a blood vessel in an image of the human retina.
Explain Hough Transform
Hough Transform is a mathematical method that can be used to detect lines, circles or other geometric shapes in an image. This method works by converting the image into parameter space, where each point in the image is represented by a set of parameters, such as the slope and the intercept of a line, or the centre and radius of a circle. The Hough Transform then searches the parameter space for clusters of points that represent lines or shapes in the image, and returns the corresponding parameters as the output.
Explain, in detail, Unsharp Masking. In each case, outline a specific example of how each of the image processing and analysis operations can be used in practice.
Used to enhance the contrast and sharpness of an image. Works by subtracting a blurred version of the image from the original image, resulting in a new image that emphasises the edges and other high-frequency details of the original image.
e.g. to improve quality of an image of a fingerprint for use in biometric identification
Explain, in detail, Intensity Normalisation. In each case, outline a specific example of how each of the image processing and analysis operations can be used in practice.
Used to adjust the brightness and contrast of an image so that its intensity values are distributed evenly across the full range of possible values. Can be achieved by mapping the intensity values of an image into a standard distribution, such as a Gaussian distribution.
This can be useful when comparing images taken under different lighting conditions or when combining multiple images into a single composite image.
Explain, in detail, DOLP filtering. In each case, outline a specific example of how each of the image processing and analysis operations can be used in practice.
Difference of low pass filtering is used to enhance the contrast of an image by suppressing the low-frequency components of the image and emphasising the high-frequency components. Can be achieved by convolving the image with a special filter kernel designed to suppress the low-frequency components of the image and enhance the high-frequency components.
Can be useful in medical imaging, where DOLP filtering can be used to improve the visibility of fine details in an image of a tissue sample.
State the difference between a Flat and Non-Flat structuring element.
Main difference: The presence or absence of internal structure. Flat SE’s are simple and easy to use, but are limited in their capabilities, while non-flat SE’s are more complex and versatile, but require more care and attention in their use.
