Abstract

Phase contrast images of stromal region of different stages of cervical pre-cancer were captured from tissue sections. A wavelet leader based multifractal analysis was performed on the phase contrast images to estimate multifractal spectrum for each image. Wavelet leaders were calculated through discrete wavelet transform using bi-orthogonal mother wavelets. The derived multifractal parameters, namely, width of singularity spectrum shows good discrimination between different grades of cervical pre-cancer.

© 2019 SPIE/OSA