Identifying Paralysis of the Sixth Optical Nerve in Video Through Convolutional Neural Networks and Curvature Isophotes

• Polyana B. Costa UFMA
• João Dallyson S. Almeida UFMA
• Jorge Antônio Meireles Teixeira UFMA
• Geraldo Braz Júnior UFMA
• Anselmo Cardoso de Paiva UFMA

DOI: https://doi.org/10.5753/sbcas.2019.6258

Abstract

This article describes a methodology to aid in the identification of the sixth nerve palsy using videos. Since the usual techniques for diagnosing this paralysis are invasive or expensive, a non-invasive and affordable method could be helpful, supporting the ophthalmologist diagnosis. The proposed method uses the convolutional neural network YOLO to detect the eyes in a face and the Isophotes Curvature to find the center of the eye. Both techniques are applied in each frame of a video in which the patient moves its eyes, keeping track of the eye movement in order to calculate the average speed in which each eye moves. The obtained results showed that the eyes that have the paralysis move 42% slower than a healthy eye.

References

Curi, R. L. N., de Oliveira Costa, I. C. B., and Barroso, T. G. M. (2013). Paralisia do vi nervo (abducente). Rev Bras Oftalmol, 72(1):59–69.

Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., and Fei-Fei, L. (2009). Imagenet: A large-scale hierarchical image database.

Ehrenhau, M. P. (2018). Abducens nerve palsy (sixth cranial nerve palsy).
GARD (2015). Sixth nerve palsy.

Gonzalez, R. C. and Woods, R. E. (2007). Image processing. Digital image processing, 2.

Gray, H. and Goss, C. M. (1974). Anatomy of the human body. American Journal of Physical Medicine & Rehabilitation, 53(6):293.

Martinsen, O. G. and Grimnes, S. (2011). Bioimpedance and bioelectricity basics. Academic press.

Ouabida, E., Essadique, A., and Bouzid, A. (2017). Vander lugt correlator based active contours for iris segmentation and tracking. Expert Systems with Applications, 71:383–395. http://dx.doi.org/10.1016/j.eswa.2016.12.001

Powers, D. M. (2011). Evaluation: from precision, recall and f-measure to roc, informedness, markedness and correlation.

Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. (2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 779–788.

Redmon, J. and Farhadi, A. (2018). Yolov3: An incremental improvement. arXiv preprint arXiv:1804.02767.

Resnick, R., Halliday, D., and WALKER, J. (2001). Fundamentos de fı́sica. Grupo Patria Cultural SA DE CV.

Rosebrock, A. (2016). Intersection over union (iou) for object detection.

Rubin, M. and Cornell, W. (2018). Paralisia do sexto nervo craniano (nervo motor ocular externo).

Szegedy, C., Ioffe, S., Vanhoucke, V., and Alemi, A. A. (2017). Inception-v4, inception-resnet and the impact of residual connections on learning. In AAAI, volume 4, page 12.

Trade, P. (2018). A keras implementation of yolov3.

Valenti, R. and Gevers, T. (2008). Accurate eye center location and tracking using isophote curvature. http://dx.doi.org/10.1109/CVPR.2008.4587529