Related Experiment Videos
Filtering noise from images with wavelet transforms
J B Weaver1, Y S Xu, D M Healy
1Department of Radiology, Dartmough-Hitchcock Medical Center, Hanover, New Hampshire.
Magnetic Resonance in Medicine
|October 11, 1991
Summary
A novel filtering technique for magnetic resonance (MR) images utilizes wavelet transforms, preserving edge sharpness while removing noise. This method may eliminate small structures comparable in size to the noise.
Area of Science:
- Medical imaging
- Signal processing
Background:
- Fourier transforms are commonly used for filtering magnetic resonance (MR) images.
- Traditional filtering methods can reduce image sharpness or fail to remove noise effectively.
Purpose of the Study:
- To introduce a new wavelet transform-based filtering method for MR images.
- To evaluate the effectiveness of this new method in noise reduction and edge preservation.
Main Methods:
- A novel image filtering approach employing wavelet transforms was developed.
- The method was applied to MR images to assess its performance.
Main Results:
- The wavelet transform-based filter successfully eliminated noise from MR images.
- Unlike Fourier transform methods, this new technique preserved the sharpness of image edges.
- A potential drawback is the elimination of small image structures similar in size to the noise.
Conclusions:
- Wavelet transform filtering offers a promising alternative for MR image processing.
- The method effectively reduces noise without compromising edge detail.
- Further research into extensions of this filtering technique is warranted.