HCRA: a hand configuration recognition algorithm for use in breast self-examination

Breast Self-examination (BSE) is a potentially life-saving skill. It is non-invasive, self-administered and a simple screening procedure for detecting breast cancer at its early stages of development. Using the hand, with a specific configuration to optimise feeling of the breast, is the basic means of performing BSE effectively. In this paper we describe an intelligent automated algorithm, for recognising the configuration of a hand that is desirable for performing the procedure. This algorithm consists of ´sobel´ detection, morphological operations and a further new method developed within the group that attempts to identify pixels delineating the edge of the hand. To demonstrate the robustness of this algorithm, the degrees-of-freedom of a normal hand are simplified, a hand class for the BSE is constructed and the system tested with 360 hand images which cover European, Asian and African ethnicities. A recognition accuracy rate of over 95% has been achieved.