Automatic Glaucoma Diagnosis with mRMR-based Feature Selection

Glaucoma\'s irreversibility, lacking of glaucoma specialists and patient unawareness demand for an economic and effective glaucoma diagnosis system for screening. In this study we explore feature selection (FS) technologies to identify the most essential parameters for automatic glaucoma diagnosis. Methods: We compose feature space from heterogeneous data sources, i.e., retinal image and eye screening data. A feature selection framework is proposed by exploring multiple feature ranking schemes and a wide range of supervised learners. The optimal feature set is derived automatically from the performance curve smoothed by measurement score regression. Results: Under the proposed framework, the optimal feature set obtained using mRMR (minimum Redundancy Maximum Relevance) scheme contains only 1/4 of the original features. The classifiers trained upon the optimal feature set are more efficient with better performance in terms of Accuracy and F-score. A detailed investigation on the features in the optimal set indicates that they can be the essential parameters for glaucoma mass screening and image segmentation. Conclusions: An intelligent Computer-aid-diagnosis (CAD) model is constructed for automatic disease diagnosis. The effectiveness of the model is demonstrated in our glaucoma study based on heterogeneous data sets. The effort not only improves the discriminative power, but also facilitates a better understanding of CAD process and may ease the data collection in glaucoma mass screening.