Numerically, what is the magnitude of the output pixel supposed to represent? I understand that the bilateral filter is an image processing technique where pixels are more "smoothed" out using other pixels of similar intensity while preserving edges -- so if the output pixel value is higher, does this mean that the pixel at the specified (x,y) location has a more sharper appearance? Or does it mean something else entirely?
nickjiang2378
I'm not totally clear how the bilateral filter is edge preserving. Wouldn't the first term mean that pixels with greater changes in intensity affect the result more?
brianqch
Based on the previous image on Bilateral Filtering, what parameters and how can we adjust to find a good balance between smoothing out noisy parts while maintaining edges?
hlxr89
The lecture mentioned that the function f takes in the difference of the intensity of two pixels, yet the equation is said to difference of pixel values. Are pixel value and intensity the same? I am assuming pixel value is the RGBa value and intensity is like brightness. Similarly for the Median Filter, Professor said that we are taking the median of the intensities of surrounding pixels, but then how does taking median of brightness result in blurring?
kalebdawit
I would imagine that this convolution equation structure that produces edge-preserving filtering could be used for other applications by modifying the f function. Are there any examples of this that are commonly used, where a filter uses the difference in intensity values from the output point for a different purpose? For instance, a simple example would be inverting the edge-preserving f function to produce an edge-selective blurring filter.
yykkcc
The bilateral filter has several key parameters like the standard deviation of the Gaussian function in the spatial domain and the value domain, which need to be carefully selected to adapt to different image and noise conditions. But I'm a little curious whether the variance of these parameters will be very large, and whether manual choosing of parameters can get the optimal solution.
Numerically, what is the magnitude of the output pixel supposed to represent? I understand that the bilateral filter is an image processing technique where pixels are more "smoothed" out using other pixels of similar intensity while preserving edges -- so if the output pixel value is higher, does this mean that the pixel at the specified (x,y) location has a more sharper appearance? Or does it mean something else entirely?
I'm not totally clear how the bilateral filter is edge preserving. Wouldn't the first term mean that pixels with greater changes in intensity affect the result more?
Based on the previous image on Bilateral Filtering, what parameters and how can we adjust to find a good balance between smoothing out noisy parts while maintaining edges?
The lecture mentioned that the function f takes in the difference of the intensity of two pixels, yet the equation is said to difference of pixel values. Are pixel value and intensity the same? I am assuming pixel value is the RGBa value and intensity is like brightness. Similarly for the Median Filter, Professor said that we are taking the median of the intensities of surrounding pixels, but then how does taking median of brightness result in blurring?
I would imagine that this convolution equation structure that produces edge-preserving filtering could be used for other applications by modifying the f function. Are there any examples of this that are commonly used, where a filter uses the difference in intensity values from the output point for a different purpose? For instance, a simple example would be inverting the edge-preserving f function to produce an edge-selective blurring filter.
The bilateral filter has several key parameters like the standard deviation of the Gaussian function in the spatial domain and the value domain, which need to be carefully selected to adapt to different image and noise conditions. But I'm a little curious whether the variance of these parameters will be very large, and whether manual choosing of parameters can get the optimal solution.