Optimally Training a Cascade Classifier

TL;DR

This paper introduces a new boosting algorithm tailored for cascade classifiers in object detection, explicitly optimizing for high detection rates and moderate false positives, leading to improved detection performance.

Contribution

It presents a principled feature selection method and a totally-corrective boosting algorithm that directly optimize the asymmetric node learning objective in cascade classifiers.

Findings

Outperforms current state-of-the-art in object detection tasks.

Effectively balances detection rate and false positive rate.

Demonstrates improved classifier performance through experimental validation.

Abstract

Cascade classifiers are widely used in real-time object detection. Different from conventional classifiers that are designed for a low overall classification error rate, a classifier in each node of the cascade is required to achieve an extremely high detection rate and moderate false positive rate. Although there are a few reported methods addressing this requirement in the context of object detection, there is no a principled feature selection method that explicitly takes into account this asymmetric node learning objective. We provide such an algorithm here. We show a special case of the biased minimax probability machine has the same formulation as the linear asymmetric classifier (LAC) of \cite{wu2005linear}. We then design a new boosting algorithm that directly optimizes the cost function of LAC. The resulting totally-corrective boosting algorithm is implemented by the column generation technique in convex optimization. Experimental results on object detection verify the effectiveness of the proposed boosting algorithm as a node classifier in cascade object detection, and show performance better than that of the current state-of-the-art.