Multiple Classifier Combination: Lessons and Next Steps

During the 1990's many methods were proposed for combining multiple classifiers for a single recognition task. With these methods, the focus of the field shifted from the competition among specific statistical, syntactic, or structural approaches to the integration of all these as potentially contributing components in a combined system. Deeper explorations of the combination methods revealed many links back to the several fundamental issues of pattern recognition. Amid the excitement and confusion, there is a persistent uncertainty in the optimality of match between a method and a problem due to strong data dependences of performance. In this paper I review several different motivations that have driven this development, summarize lessons learned in the exploration of combination methods, outline the difficulties encountered, and suggest ways to break out of the current plateau.