Hamiltonian Streamline Guided Feature Extraction with Applications to Face Detection

TL;DR

This paper introduces a novel feature extraction method using Hamiltonian streamlines derived from intensity landscapes, which provides efficient and topologically rich features for object detection, notably face detection.

Contribution

The method leverages Hamiltonian systems for feature extraction, resulting in smaller feature spaces, faster training, and comparable detection performance to Haar-like features.

Findings

Feature space is significantly smaller than Haar-like features.

Training time is extremely short.

Detection speed and accuracy are comparable to Haar-like based classifiers.

Abstract

We propose a new feature extraction method based on two dynamical systems induced by intensity landscape: the negative gradient system and the Hamiltonian system. We build features based on the Hamiltonian streamlines. These features contain nice global topological information about the intensity landscape, and can be used for object detection. We show that for training images of same size, our feature space is much smaller than that generated by Haar-like features. The training time is extremely short, and detection speed and accuracy is similar to Haar-like feature based classifiers.