Deep Learning Provides High Accuracy in Automated Chondrocyte Viability Assessment in Articular Cartilage Using Nonlinear Optical Microscopy

Chondrocyte viability is a crucial factor in evaluating cartilage health. Most prevalent cell viability assays rely on dyes and are not applicable for in vivo or longitudinal studies. We previously demonstrated that the two-photon excited autofluorescence and second harmonic generation microscopy provided high-resolution images of cells and the collagen structure; those images allowed us to distinguish live/dead chondrocytes by visual assessment or by the normalized autofluorescence ratio. However, both methods rely on human involvement and their throughputs are low. Methods for automated cell-based image processing are desired to improve the analysis throughput. Conventional image processing algorithms do not perform well on autofluorescence images acquired by nonlinear microscopes due to the low image contrast. In this study, we compared conventional, machine learning, and deep learning methods in chondrocyte segmentation and classification. We demonstrated that deep learning significantly improved the outcome of the chondrocyte segmentation and classification; with appropriate training, the deep learning method can achieve the accuracy of 90% in chondrocyte viability measurement. The significance of this work is that the automated imaging analysis is possible and should not become a major hurdle in putting nonlinear optical imaging methods to practical uses in biological or clinical studies.