Image quality assessment for performance evaluation of despeckle filters in Optical Coherence Tomography of human skin

Abstract

The objective of this work is to evaluate the performance of a set of despeckle filters for Optical Coherence Tomography (OCT) of the skin. The six filters are based on local statistics, median filtering, pixel homogeneity, geometric filtering and transformed domain homomorphic filtering. The results of this study suggest that geometric filtering algorithm outperforms other candidate methods, and exhibits minimum Minkowski metric. It also has SNR > 50dB, with best image quality and intact structural similarity. The results of image quality are also related with visual assessment by an expert. The experiments performed here for selection of a proper despeckle filter for OCT requires further large scale evaluation for application in clinical practice, exploratory analysis, automated segmentation, texture analysis, and image based classification techniques. On a wider spectrum this work is also useful as a framework for comparative assessment of similar image quality improvement algorithms.