Transitional appearance-based motion tracking for real-time breast self-examination supervision.

Despite much research and clinical trials, breast cancer still presents a serious threat of premature death to women. Early detection of the lumps in the breast is a key contributing factor to the successful treatment of this potentially fatal disease. Performing breast self-examination (BSE) in an accurate manner can assist a woman in detecting any abnormalities in her breasts, which may mark the onset of potential disease. This is also an essential tool used to enhance breast awareness. Using the hand, in a specific configuration, and palpating the entire breast in a certain movement pattern can optimise her feeling of the breast, In this paper we describe an intelligent automated algorithm for tracking the finger pads of a moving hand with the movement videos captured by a common web camera. The algorithm employs the principle of HCRA (hand configuration recognition algorithm) and its refinement through a simple but novel transitional appearance-based model. A novel hand motion recognition algorithm (HMRA) is developed to recognise the motion pattern. Desirable tracking and recognition results have been achieved and the robustness of this algorithm is demonstrated in this paper.