Summary: | 碩士 === 國立中興大學 === 應用數學系所 === 106 === New detection technology of ophthalmic instrument brings more convenience and accuracy on treatment to doctors. For example, the improved ophthalmoscope was invented, and the optical coherence tomography was introduced in ophthalmology. It all helps the ophthalmologists diagnose the eye disease more accurately. As a result, this article aims at establishing a 3D model for eye OCT images and calculating subretinal fluid volume, in that way, we can help doctors to determine the patient''s recovery condition and estimate the recovery time accurately.
Subretinal fluid contour is a closed curve, so we recommend take Balloon Snake as the main algorithm to detect subretinal fluid contour. We first adopt Gaussian filter and brightness adjustment formula to do preprocessing, using cubic spline interpolation and fuzzy function to build the required contours’ mathematical model, finally establish 3D model of subretinal fluid and calculate the volume.
After 20 medical cases, we make the comparison with the calculated volume and the volume estimated by the contour drawn by professional staff, the posterior intraclass correlation coefficient is up to 0.9934. It provides fast and stable algorithms in the calculation of subretinal fluid’s volume and the establishment of the 3D model.
|