Study Of Motion Artifacts Correction Algorithm In Optical Coherence Tomography Images

Optical coherence tomography (OCT) is a novel and noninvasive optical tomography technique in modern biomedical imaging field, which can obtain 2D/3D image of the internal microstructure in biological systems, and which has many advantages,such as high resolution, high sensitivity, noninvasive, in vivo real-time imaging,depth and transverse resolutions are independent and so on. OCT has been the research focus of biomedical noninvasive imaging since it was first introduced; it has become a standard tool for the diagnosis and study of eye disease. OCT has made its significant clinical contribution in the field of ophthalmology and become the "gold standard" for early diagnosis of retinal disease and glaucoma.In the OCT dynamic biopsy, motion artifact can easily be generated by head, eye and other uncontrollable factors motion. Motion artifact will make the volume data sets loss authenticity; structure information will be lost in three-dimensional images, so that the physical structure of the retina can not get the real reaction. The high sensitivity in OCT real-time imaging even leads to imaging failure, and may reduce the practicability of OCT in clinical. Therefore, motion artifacts correction has become an urgent issue in optical coherence tomography imaging.In this paper, an algorithm was designed to correct the distortion volume data set. Motion artifact in X and Y directions of volume data sets have been corrected respectively without any hardware conditions. First, to reduce the noise with the improved complex nonlinear diffusion filtering as preprocessing, then to roughly correct with the center of image, finally, to correct motion artifacts along X direction with normalized cross-correlation method and to correct motion artifacts along Y direction with the difference between average method. The results show that after correction, the motion artifacts in the longitudinal 2D images and the 3D image disappeared, the surface of the 2D and the 3D image became more smooth, the structure between layers of the images got clear and distinct, retinal en-face single slice image was sharp,and the fundus tissue structure could be observed. Conclusion: The algorithm of correction makes the physical structure of the retinal display truly. It may reduce the risk of misdiagnosis and improve the accuracy and objectivity in clinical diagnosis.