Fast features for time constrained object detection

This paper concerns itself with the development and design of fast features suitable for time constrained object detection. Primarily we consider three aspects of feature design; the form of the precomputed datatype (e.g. the integral image), the form of the features themselves (i.e. the measurements made of an image), and the models/weak- learners used to construct weak classifiers (class, non-class statistics). The paper is laid out as a guide to feature designers, demonstrating how appropriate choices in combining the above three characteristics can prevent bottlenecks in the run-time evaluation of classifiers. This leads to reductions in the computational time of the features themselves and, by providing more discriminant features, reductions in the time taken to reach specific classification error rates. Results are compared using variants of the well known Haar-like feature types, Rectangular Histogram of Oriented Gradient (RHOG) features and a special set of Histogram of Oriented Gradient features which are highly optimized for speed. Experimental results suggest the adoption of this set of features for time-critical applications. Time-constrained comparisons are presented using pedestrian and road sign detection problems. Comparison results are presented on time-error plots, which are a replacement of the traditional ROC performance curves.