Abstract
Purpose: :
Our aim is to understand the nature of visual processing in retina dystrophy patients and hence developing an augmented vision system which aids key feature recognition for them. This study assesses which scene enhancements, if any, provide benefit in terms of enhanced recognition ability and potential quality of life on groups of patients with macular degenerations. We will also describe our ongoing effort to combine this system into a wearable augmented vision headset
Methods: :
Based on preliminary model of degenerate retina, we developed custom processing algorithms including Cartoonization (spatial simplification with enhanced edge), Booster (a weighted color image based on edges strength), Edge overlay (on cartoonized or original images) and Motion enhancements. Testing has been carried out at Liaquat University Eye Hospital, Pakistan (200 Patient study) and Oxford Eye Hospital, UK (45 patient study). At the time of writing 33 have been tested in Liaquat and 6 patients have been tested at Oxford. Each centre had separate but complementary tests. These consist of showing the patients sets of images and assessing determination ability, determination speed and preference. In liaquat the tests concentrated on basic shapes and edges. In Oxford a broader test was performed examining more complex shapes, features, and algorithms.
Results: :
Analysis demonstrated significant improvement in patients’ ability to recognize major image features with image processing. Outcome varied with image type. Cartoonization was most preferable for images with low contrast, luminance and feature size and for videos with complex features and complex motion. Edges overlay, on the other hand, were preferred for scenes with high luminance, high contrast and large-major-features as well as videos with less complex motion and scene detail.
Conclusions: :
Comparing all images and videos, Cartoonization was picked more frequently than other algorithms for aiding recognition of features and motion and willingness to use in daily life.
Keywords: image processing • age-related macular degeneration