Image software identifies people by their ears

Researchers at Southampton University have developed a new image software program that can be used to identify people by their ears.

The so-called ’Image Ray Transform’ can be used to extract tubular and circular features in an image − such as the curved rim at the top of the ear, known as the helix − that are not often found by other techniques.

The transform was developed by Prof Mark Nixon, Dr John Carter and Alastair Cummings from the university’s School of Electronics and Computer Science. Prof Nixon, one of the UK’s earliest researchers in this field, first proved that ears were a viable biometric in 2005.

Cummings explained that the Image Ray Transform works by acquiring an image of the ear and uploading it to a computer. The program then shines thousands of virtual light rays through the image.

The paths of these rays are used to generate a new image, with tubular and circular features emphasised. The image is then reformatted into a matrix of two-dimensional glass blocks, each representing a pixel whose refractive index is related to the intensity of the pixel in the image.

Cummings said tubular features such as the helix are represented by bands of slightly changing intensity.

Tests showed that the new technique achieved 99.6 per cent success at identifying ears from over 250 images, despite hair concealment and possible confusion with spectacles.

Cummings said ear identification through the Image Ray Transform technique could be used with facial recognition and iris scans to give improved biometric detection.

He also imagined that the technique could eventually be used to biometrically identify a criminal on CCTV as the ear − unlike an iris scan, for example − can be evaluated from images taken further away.