HEARD: An automatic human EAR detection technique

A new class of biometrics based upon ear features was introduced for use in the development of passive identification. This paper presents an efficient technique for automatic ear detection from side face images. The proposed technique detects human's ear without any training or assumption of prior knowledge about the input image, and it does not need any human intervention so it can be used successfully in a fully automated ear recognition system. Two new features have been proposed for ear detection: the first represents elongation and is expressed by the ratio between the boundary's width and height. The second describes compactness and is measured by the ratio between the number of pixels composing the boundary's area and the number of pixels composing its perimeter. Each ratio should exceed a certain threshold for human's ear. Moreover, the proposed approach relies on the fact that the ear is a semi rounded boundary which contains other smaller semi rounded boundaries. The proposed technique uses the roundness as another feature. In human's ear, roundness falls within a certain range. Using these three features better detection performance is achieved and less time was needed. The experimental results demonstrate the effectiveness of the approach.