Have you ever gotten your pictures back from the developers only to be disappointed by the demonic red eyes your family now displays? Have you wondered why the red-eye function on your camera never works, or how you could get rid of those red-eyes from your favourite picture?

At U of T, Professor Konstantinos Plataniotis of the department of electrical and computer engineering and his colleagues have developed a new technology that would allow consumers to remove the red-eye from photographs using their digital cameras. In fact, with the technology they hope to have in place within the year, the camera would detect and remove the red-eye for you.

Red-eye occurs in photographs when the blood in the retina absorbs all colours from the flash except for red. As Plataniotis explained, existing technologies try to correct red-eye after a picture has been processed, either using software to remove the red, or by manually using a black pen. The other method is a series of timed flashes that go off before the picture is taken. This often reduces red-eye, but does not completely rectify it.

The new technology would be built right into a camera, and works in three phases. The first component processes skin tone in order to gauge the colour of the subject’s skin, aiding the camera in deciding, based on population norms, what colour the eyes should be (e.g. dark hair and dark eyes go together). The second is a shape-adaptive module that locates the retina on the subject’s face. The final step is a pixel processor that performs calculations using an algorithm to adjust the colour of the eye from red to the predetermined natural colour. These three step combines correct the red-eye before the picture is displayed to the consumer.

This technology is the result of a collaboration between two continents and three different countries, with collaborators in Poland and Norway. The initial research was conducted at U of T in 1995 by a student writing her undergraduate thesis. Without her initial effort, Plataniotis said much of this work would not have reached the stage it has. This technology could also be applicable as a safety tool to monitor the eyes of truck drivers for fatigue; it would watch the shape of the eye and iris and alert the driver when their eyes were getting tired.