This project was created to investigate noise-reduction algorithms for a future project that involves building a deep learning model to detect diseases in medical imagery. Using a sliding window approach, the following linear and non-linear noise-reduction algorithms were implemented and compared:
- Mean filtering (fmean)
- Median filtering (fmedian)
- Sharpening (fsharpen)
- Unsharp masking (funsharp)
- Gaussian filtering (fgaussian)
- Alpha-trimmed mean filtering (ftrimmed)
- Adaptive weighted median filtering (fadaptive)
- Truncated median filtering (ftruncated)
- Morphological reconstruction (fmorph)
- Kuwahara filtering (fkuwahara)