Investigation of event-based memory surfaces for high-speed tracking, unsupervised feature extraction and object recognition

classifying model airplanes that are dropped free-hand close to the camera lens so as to generate a challenging highly varied dataset of spatio-temporal event patterns. We investigate the use of time decaying memory surfaces to capture the temporal aspect of the event-based data. These surfaces are then used to perform unsupervised feature extraction, tracking and recognition. Both linear and exponentially decaying surfaces were found to result in equally high recognition accuracy. Using only twenty five event-based feature extracting neurons in series with a linear classifier, the system achieves 98.61% recognition accuracy within 156 milliseconds of the airplane entering the field of view. By comparing the linear classifier results to a high-capacity ELM classifier, we find that a small number of event-based feature extractors can effectively project the complex spatio-temporal event patterns of the data-set to a linearly separable representation in the feature space.