Novel Fusion Methods for Pattern Recognition
Muhammad Awais, Fei Yan, Krystian Mikolajczyk and Josef Kittler
In: The European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 5 - 9 September 2011, Athens, Greece.
Over the last few years, several approaches have been proposed for information fusion including dierent variants of classier level fusion (ensemble methods), stacking and multiple kernel learning (MKL). MKL has become a preferred choice for information fusion in object recognition. However, in the case of highly discriminative and complementary feature channels, it does not signicantly improve upon its trivial baseline which averages the kernels. Alternative ways are stacking and classier level fusion (CLF) which rely on a two phase approach. There is a signicant amount of work on linear programming formulations of ensemble methods particularly in the case of binary classication.
In this paper we propose a multiclass extension of binary nu-LPBoost, which learns the contribution of each class in each feature channel. The existing approaches of classier fusion promote sparse features combinations, due to regularization based on l1-norm, and lead to a selection of a subset of feature channels, which is not good in the case of informative channels. Therefore, we generalize existing classier fusion formulations to arbitrary lp-norm for binary and multiclass problems which results in more eective use of complementary information. We also extended stacking for both binary and multiclass datasets.We present an extensive evaluation of the fusion methods on four datasets involving kernels that are all informative and achieve state-of-the-art results on all of them.