Computer-aided BSE torso tracking algorithm using neural networks, contours, and edge features
Rey Anthony A. Masilang, Melvin K. Cabatuan, Elmer P. Dadios, Laurence A. Gan Lim · 2014
This paper presents an algorithm for tracking the torso of the user in a computer-aided breast self-examination system. The algorithm uses a neural network-based skin classifier for segmenting the skin area from the non-skin area. Using the skin mask produced by the classifier, the contours of the body are extracted and used to identify the region containing the torso of the user. The algorithm is tested on 4 different videos. The performance of the algorithm is measured in terms of its F1-score. Results show that the algorithm is capable of accurate tracking with an F1-score of 92.97%.