کنفرانس بین المللی محاسبات نرم

صفحه اصلی / چهارمین کنفرانس بین المللی محاسبات نرم

3D intra-retinal layer segmentation of OCT data using modified live wire algorithm

نویسندگان :

Hamed Fahimi¹ Jalal Chachi²

1- دانشگاه شهید چمران اهواز 2- دانشگاه شهید چمران اهواز

کلمات کلیدی :

Optical Coherence Tomography, Image segmentation, automatic method, interactive method, 2D boundary, active learning

چکیده :

Image segmentation is the process of dividing a digital image into sets of pixels. Its approaches are two types: automated and interactive. In contrast to automated methods, interactive algorithms, such as live wire with a core in graph theory, do not need user assistance and provide complete control to the user, hence they are a desirable approach in many medical applications. Particularly, optical coherence tomography (OCT) is a powerful method for imaging the retinal layers. Towards attaining a 3D version of live wire algorithm, which is not as simple as the 2D case, this paper makes a step forward. We develop existing versions of this algorithms. We experiment over OCT data of 13 eye patients and demonstrate how they improve image segmentation of different retinal layers. Moreover, we show the good effect of two different algorithms of noise reduction on OCT images on the performance of modified live wire algorithm.