Abstract
Amblyopia or "lazy" eye is a disorder of the visual system that causes poor vision in an eye that is otherwise physically
normal, and it affects 2-3% of the population, which equates to conservatively around 10 million people under the age of 8
years worldwide. Amblyopia is a neurologically active process: the problem is caused by either no transmission or poor transmission
of the visual stimulation through the optic nerve to the brain. With time, if no treatmeant is performed, the weak eye becomes
even weaker and the other eye becomes dominant. Amblyopia is classically treated by clarifying the visual image with glasses,
and patching (totally or partially) the dominant eye in order to force the use of the amblyopic eye. Patching suffers from
several problems: it is unpopular, prolonged, and it can sometimes disrupts any residual fusion between the visions of the
eyes. This results often in noncompliance with the therapy. Several alternatives have been introduced, including partial occlusion
and vision rebalancing in which the image to the lazy eye is enhanced and the image to the good eye is penalized. We present
how a 3D technology can be used to realize a system for vision rebalancing of video clips which exploits the stereo vision
of the 3D system. This technology is relatively inexpensive, easy to use also in a domestic environment, with recrational
activities enjoyable by the children, and easy to extend. We have implemented a prototype software system which processes
a video and sends a penalized version to the good eye and an enhanced version to the lazy eye. We use a framesever for runtime
video processing and several image filters and meta-filters to obtain the final video to be viewed by the patient. We argue
for the viability of the proposed method in the treatment of amblyopic children.
[download the pdf file] [DOI] [presentation]