Barbara Caputo
This paper presents a novel SVM-based algorithm for visual object recognition, capable of learning model representations incrementally. We combine an incremental extension of SVMs with a method which reduces the number of support vectors needed to build th ...
2006