Real-Time Reservoir Computing Network-Based Systems for Detection Tasks on Visual Contents

Among the various types of artificial neural networks used for event detection in visual contents, those with the ability of processing temporal information, such as recurrent neural networks, have been proved to be more effective. However, training of such networks is often difficult and time consuming. In this work, we show how Reservoir Computing Networks (RCNs) can be used for detecting purposes on raw images. The applicability of RCNs is illustrated using two example challenges, namely isolated digit handwriting recognition on the MNIST dataset as well as detection of the status of a door using self-developed moving pictures from a surveillance camera. Achieving an error rate of 0.92 percent on MNIST, we show that RCN can be a serious competitor to the state-of-the-art. Moreover, we show how RCNs with their simple and yet robust training procedure can be practically used for real surveillance tasks using very low resolution camera sensors.

I. Introduction

Many types of Artificial Neural Networks (ANN) have been used through time to improve the performance of object recognizers. Multilayer Perceptrons (MLPs) [1] were among the first classifiers that were tested for isolated handwritten digit recognition [2]. MLPs have been the subject of many research efforts because of their interesting learning and generalization properties. However, a drawback of MLPs is the inability to process temporal information, which is vital to process observations that change slowly over time. A solution to this problem could be to take a window of frames in the network input instead of only one. Another approach is to add recurrent connections to the network, leading to Recurrent Neural Networks (RNNs) [3].