Research On Detection Method Of Retinal Nerve Fiber Layer Pulse Based On Optical Coherence Tomography

Glaucoma is an eye disease with a high rate of blindness,early diagnosis and treatment can significantly reduce injury.The thickness of retinal nerve fiber layer is related to the development of glaucoma course,pulse amplitude of retinal nerve fiber layer associated with vascular disease.These two parameters are important to evaluate the biomechanical characteristics of eye tissue and provide important basis for early diagnosis of glaucoma.Because of the influence of body movement,the extraction of human eye RNFL pulsation has not been realized.Spectral domain optical coherence tomography(SD-OCT)can achieve fast high-resolution imaging and obtain phase changes of small motion interference signals.It has a high prospect in extracting RNFL in body beating signals.To meet the needs of early diagnosis of glaucoma,the SD-OCT pulse parameter extraction based on image segmentation,phase extraction and body motion compensation is studied in this paper.A system and algorithm for extracting human eye RNFL pulsation are validated in this paper:On the basis of SD-OCT instruments,an experimental method was designed to locate RNFL large vessels near optic papilla by line scanning laser ophthalmoscope,and the RNFL data were collected in different states of healthy people.We propose a RNFL segmentation algorithm based on full convolutional neural networks.After dividing the collected data using the shortest path method as a standard training convolutional neural network.The trained network can remove other parts affecting RNFL phase extraction and improve the accuracy of the phase extraction algorithm on RNFL pulsation extraction.A histogram-based RNFL pulsation extraction algorithm is proposed in this paper.This algorithm tracks the displacement of the RNFL by extracting phase difference changes between adjacent images.The algorithm uses image registration and linear fitting motion model to evaluate the moving noise phase of tissue,eliminate the moving phase of body,and realize the extraction of pulsating phase.An infrared pulse sensor device is designed to collect pulse.At the same time,pulse and OCT data were collected.Finally,the pulse waveform and the RNFL pulse waveform are compared and analyzed in time domain and frequency domain.RNFL pulsation extraction method may provide valuable parameters for early diagnosis of glaucoma.It has important research significance and application value for early diagnosis of glaucoma.