Abstract

For noise reduction in optical coherence tomography (OCT) image, wavelet domain thresholding has the unique advantage of suppressing noise while preserving image sharpness. However, previous applications of wavelet domain thresholding did not consider the fact that speckle noise has different characteristics in different spatial scales. In this study, we demonstrate a noise adaptive wavelet thresholding (NAWT) technology that exploits the difference in noise characteristics in different wavelet sub-bands. Our results demonstrated that NAWT outperforms conventional wavelet thresholding.

© 2017 Optical Society of America