A Conditional Random Field Weakly Supervised Segmentation Approach for Segmenting keratocytes Cells in Corneal Optical Coherence Tomography Images

Abstract

Keratocytes are vital for maintaining the overall health of human
cornea as they preserve the corneal transparency and help in healing
corneal injuries. Manual segmentation of keratocytes is challenging,
time consuming and also needs an expert. Here, we propose
a novel semi-automatic segmentation framework, called Conditional
Random FieldWeakly Supervised Segmentation (CRF-WSS)
to perform the keratocytes cell segmentation. The proposed framework
exploits the concept of dictionary learning in a sparse model
along with the Conditional Random Field (CRF) modeling to segment
keratocytes cells in Ultra High Resolution Optical Coherence
Tomography (UHR-OCT) images of human cornea. The results
show higher accuracy for the proposed CRF-WSS framework compare
to the other tested Supervised Segmentation (SS) andWeakly
Supervised Segmentation (WSS) methods.