Digital Post Processing - Part 2 Flashcards
(45 cards)
What is Detail Processing?
Processing that treats fine details as separate components of an image, which out changing the overall brightness/conrtast of an image
What two domains is detail processing usually done in?
Spatial domain and frequency domain
How is detail processing done in the spatial domain?
By applying kernels
What is a kernel?
A submatrix or smaller “core” matrix of the imaging which are adjusted by applying mathematical functions to it
What detail functions are done through applying kernels?
- Noise reduction
- Edge enhancement
- Background suppression
How does the process of applying kernels work?
Each pixel in the image is multipled by the value of the kernel and then summed up to form a new pixel in the area the kernel is centered
How is image noise reduced via kerneling?
If the 8 surrounding squares all have the same values, then its applied to the middle square which reduces noise
How is edge enhancment done via kerneling?
The sum of the 8 surrounding squares equals 1, which is added to the middle square which enhances the edges
What is the tradeoff to edge enhancement?
Increased noise
How is background suppression done via kerneling?
The sum of the 8 surrounding squares equals 0, which is then added to the middle square to suppress the background
What additional processing technique can be applied in the spatial domain for detail processing?
Unsharp Mask Filtering
What is Unsharp Mask Filtering?
A three step process to increase fine details
What are the steps to Unsharp mask filtering?
- Image created from the original image that only contains large gross structues (blurry image since details are removed)
- Positive mask is made of the image which inverts the contrast of each pixel
- The postitive mask is subtracted from the original image, which leaves only smaller structures
What image qualities does the new image have after applying the unsharp mask filter?
Image is noisy but has super spatial resolution
Why would an unsharp mask filter be used?
To allow a radiologist to see specific details in an image that they otherwise would not be able to see (hairline fractures)
How is detail processing done in the frequency domain?
By breaking the image down into wave components and removing certain frequencies to make adjustements
What two processing techniques can be adjusted in the frequency domain?
- Noise reduction
- Edge enhancement
How is noise reduction achieved in the frequency domain?
Shorter wavelengths (small objects) are removed from the image, which decreases spatial resolution
How is edge enhancement achieved in the frequency domain?
Longer wavelengths (large objects) are removed from the image, which increases spatial resolution
What is the relationship between object size and frequency?
- Small objects/details have short wavelengths and higher frequency
- Large objects/details have long wavelengths and lower frequencies
What is the relationship between wavelength and pixel size?
Wavelength equals pixel size
What is the relationship between aplitude and contrast?
- High amplitude waves have a higher contrast (short scale)
- Low amplutide waves have a lower contrast (long scale)
If the amplitude of a wave is changed what effect will it have on contrast?
It will alter the scale of contrast
How are details and frequency related?
Lower frequency waves have less details, since more pixels are represented in each wave.
Higher frequency waves have more detail (sharpness), since less pixels are represented in each wave
