Detecting and tracking female breasts using neural network in real-time

The general aim of this research is helping women to perform breast self-examination (BSE) for finding out any abnormality, change, or lump in the breasts. BSE involves checking the breasts for finding abnormalities, lumps, or changes. This paper reports about our initial efforts to detect and track the left and right breasts in real-time imaging. Image frames were processed considering the color information, and integral image processing to segment regions of interest (ROI) according to common colors of breast features. After getting the preliminary candidate regions, the vector of features were used as the inputs of neural network. The algorithm applies each ROI into the artificial neural network (ANN) for detection of the right and left breasts. Results of the study show that the proposed ANN successfully identifies the position and location of the breasts.