The Bitonic Filter: Linear Filtering in an Edge-Preserving Morphological Framework
read more
Citations
Hyperspectral Anomaly Detection With Attribute and Edge-Preserving Filters
Image denoising review: From classical to state-of-the-art approaches
Speckle Noise Reduction Technique for SAR Images Using Statistical Characteristics of Speckle Noise and Discrete Wavelet Transform
Despeckling Images Using a Preprocessing Filter and Discrete Wavelet Transform-Based Noise Reduction Techniques
Using Taylor Expansion and Convolutional Sparse Representation for Image Fusion
References
Image quality assessment: from error visibility to structural similarity
Scale-space and edge detection using anisotropic diffusion
Bilateral filtering for gray and color images
Guided Image Filtering
A Review of Image Denoising Algorithms, with a New One
Related Papers (5)
Frequently Asked Questions (11)
Q2. What is the filter for low-noise images?
For low-noise images, the filter can be used to smooth over repetitive details (anything not bitonic overthe filter length) whilst preserving individual fine details and any transitions very well, which is potentially of use in edge extraction, background extraction, feature enhancement or other artistic processing.
Q3. What is the meaning of the term order-statistic filters?
Since bitonicity is concerned with the ordering, rather than the value, of the data, it is natural to turn to rank filters [2], also known as order-statistic filters.
Q4. What is the main disadvantage of the bitonic filter?
Since the bitonic filter is entirely local and not iterative, adaptations of other filters (e.g. iteration or local parameter variation) might also be appropriate adaptations of the bitonic.
Q5. What is the filter parameter for smoothing in non-local means?
Smoothing in non-local means is controlled by the filter parameter h, which is usually set somewhat higher than the expected noise level in the image.
Q6. What is the earliest method of filtering?
Possibly the earliest is the alpha-trimmed mean [21], which uses a rank filter to remove outliers before taking the mean (effectively a simple linear filter) of the remainder.
Q7. What is the effect of the shape of the window on the data?
For two-dimensional (2D) data, the shape of the window used to form the set of ranked data, in morphology known as the ‘structuring element’, has some impact on which features can be preserved.
Q8. What is the difference between the bitonic and the non-local means filters?
The other linear filters are not built on a bitonic definition of signals, and are more naturally sensitive to data values, with the associated difficulty of preserving any edges in the data which are smaller than the level of the noise.
Q9. What is the reason for the poor performance of the median?
This analysisalso explains the poor performance of the median: the gain is in fact negative for some frequencies, and there is considerable distortion, particularly for low frequencies.
Q10. What is the difference between rank filters and morphological filters?
These filters vary in rank selection, but will return one of the input values: hence they are limited in their power to reduce non-impulsive noise, though more complex combinations of morphological operations such as discussed by [12] have been used to de-noise specific signals [13].
Q11. What is the equivalent of a larger weight in linear filtering?
The filter can then either be modified in extent, or some data weighted by repeating values in the ranked list: the rank filter equivalent of multiplying by a larger weight in linear filtering.