advertisement
Abstract #82099 Published in IGR 20-4
Adaptive weighted locality-constrained sparse coding for glaucoma diagnosis
Zhou W; Yi Y; Bao J; Wang WMedical and Biological Engineering and Computing 2019; 57: 2055-2067
Glaucoma is a sight-threading disease which can lead to irreversible blindness. Currently, extracting the vertical cup-to-disc ratio (CDR) from 2D retinal fundus images is promising for automatic glaucoma diagnosis. In this paper, we present a novel sparse coding approach for glaucoma diagnosis called adaptive weighted locality-constrained sparse coding (AWLCSC). Different from the existing reconstruction-based glaucoma diagnosis approaches, the weighted matrix in AWLCSC is constructed by adaptively fusing multiple distance measurement information between the reference images and the testing image, making our approach more robust and effective to glaucoma diagnosis. In our approach, the disc image is firstly extracted and reconstructed according to the proposed AWLCSC technique. Then, with the usage of the obtained reconstruction coefficients and a series of reference disc images with known CDRs, the CDR of the testing disc image can be automated estimation for glaucoma diagnosis. The performance of the proposed AWLCSC is evaluated on two publicly available DRISHTI-GS1 and RIM-ONE r2 databases. The experimental results indicate that the proposed approach outperforms the state-of-the-art approaches. Graphical abstract The flowchart of the proposed approach for glaucoma diagnosis.
College of Computer Science, Shenyang Aerospace University, Shenyang, Liaoning, China. zhouweineu@outlook.com.
Full articleClassification:
2.14 Optic disc (Part of: 2 Anatomical structures in glaucoma)
6.9.5 Other (Part of: 6 Clinical examination methods > 6.9 Computerized image analysis)
6.30 Other (Part of: 6 Clinical examination methods)