Loke, Sankrandan and Jana, Soumya
(2018)
3D Image Reconstruction of Retina using Aplanats.
Masters thesis, Indian Institute of Technology Hyderabad.
Abstract
Existing fundus cameras provide two-dimensional images for the fundus or retina which are curved
in shape, thereby distorting the peripheral information. For a quality and aberration free image, a
system of lenses need to be incorporated for imaging. But that comes with narrow field of view,
high-cost, and makes it bulkier and difficult to operate. Retinal abnormalities outside the photo-
graphic field cannot be captured with narrow field of view. Any compromise with the quality of
imaging system will lead to poor fundus image quality. In this thesis, an attempt has been made
to demonstrate that aplanats can be used to replace the lens system in fundus cameras with the
benefit of acheiving high field of views of about 200o without any pupil dialation needed. Aplanats
are aberration free and have high numerical apertures. Therefore, they can even collect highly di-
verging rays. But due to their design, complete retina cannot be imaged using a single aplanat. So
a three phase imaging procedure is proposed in which three annular regions of retina are imaged
in each phase. In the first two phases imaging is done by the aplanats and for the third stage a
simple mobile can also be used. This imaging covers the whole retina, with some overlaps so that
the images collected can be stitched together post that. In first two phases, sensors are specially
designed which capture the 3D information of retina. The 3D reconstruction of complete retina from
sensor data is discussed later in detail. A developed ray tracer is talked about which was used as the
tool to work towards the aim. Though aplanats are aberration free, the light reflected from retina
has to pass through the eye-lens which makes the aggregate system aberrant. After the 3D retina
image is constructed, all the details can be identified but with slight blur. With appropriate signal
processing techniques, a sharp 3D image of retina can be obtained.
Actions (login required)
|
View Item |