Two-stage Augmented Kernel Matrix for Object Recognition
Muhammad Awais, Fei Yan, Krystian Mikolajczyk and Josef Kittler
In: 10th International Workshop on Multiple Classifier Systems, 15-17 June 2011, Naples, Italy.
Multiple Kernel Learning (MKL) has become a preferred choice for information fusion in image recognition problem. Aim of MKL is to learn optimal combination of kernels formed from dierent features, thus, to learn importance of dierent feature spaces for classication. Augmented Kernel Matrix (AKM) has recently been proposed to accommodate for the fact that a single training example may have dierent importance in dierent feature spaces, in contrast to MKL that assigns same weight to all examples in one feature space. However, AKM approach is limited to small datasets due to its memory requirements. We propose a novel two stage technique to make AKM applicable to large data problems. In rst stage various kernels are combined into dierent groups automatically using kernel alignment. Next, most in
uential training examples are identied within each group and used to construct an AKM of signicantly reduced size. This reduced size AKM leads to same results as the original AKM. We demonstrate that proposed two stage approach is memory ecient and leads to better performance than original AKM and is robust to noise. Results are compared with other state-of-the art MKL techniques, and show improvement on challenging object recognition benchmarks.