An efficient deep learning segmentation scheme for cervical collagen and elastin quantification in Mueller matrix polarimetry microscopic images
Conference
Gary, N, Le, VND, Wojak, J et al. (2022). An efficient deep learning segmentation scheme for cervical collagen and elastin quantification in Mueller matrix polarimetry microscopic images
.
Gary, N, Le, VND, Wojak, J et al. (2022). An efficient deep learning segmentation scheme for cervical collagen and elastin quantification in Mueller matrix polarimetry microscopic images
.
We present an efficient deep learning-baseds egmentation approach able to discriminate collagen from elastic fibers from Mueller matrix microscopy images of the mouse cervix. Thanks to the use of a Self-Validating Mueller matrix Micro-Mesoscope (SAMMM) system, Second Harmonic Generation (SHG) and Two Photon Excitation Fluorescence (TPEF) are also acquired and used as the quantitative segmentation ground truth. The method combines a multilayer perceptron and two U-net convolution networks. The accuracy and image quality metrics show improved segmentation results.