Insights from Classifying Visual Concepts with Multiple Kernel Learning

TL;DR

This paper evaluates non-sparse multiple kernel learning (MKL) methods for visual concept classification, demonstrating their advantages and limitations compared to sum kernel and sparse MKL on benchmark datasets.

Contribution

It applies a recent non-sparse MKL variant to computer vision tasks and compares its performance with traditional methods, providing new insights into MKL fusion strategies.

Findings

Non-sparse MKL outperforms sum kernel in certain scenarios.

Sparse MKL often underperforms compared to unweighted sum kernel.

Empirical results on PASCAL VOC 2009 and ImageCLEF2010 datasets support these conclusions.

Abstract

Combining information from various image features has become a standard technique in concept recognition tasks. However, the optimal way of fusing the resulting kernel functions is usually unknown in practical applications. Multiple kernel learning (MKL) techniques allow to determine an optimal linear combination of such similarity matrices. Classical approaches to MKL promote sparse mixtures. Unfortunately, so-called 1-norm MKL variants are often observed to be outperformed by an unweighted sum kernel. The contribution of this paper is twofold: We apply a recently developed non-sparse MKL variant to state-of-the-art concept recognition tasks within computer vision. We provide insights on benefits and limits of non-sparse MKL and compare it against its direct competitors, the sum kernel SVM and the sparse MKL. We report empirical results for the PASCAL VOC 2009 Classification and ImageCLEF2010 Photo Annotation challenge data sets. About to be submitted to PLoS ONE.