Abstract. We propose in this paper a fully automated deep model, which learns to classify human actions without using any prior knowledge. The first step of our scheme, based on the extension of Convolutional Neural Networks to 3D, automatically learns spatio-temporal features. A Recurrent Neural Network is then trained to classify each sequence considering the temporal evolution of the learned features for each timestep. Experimental results on the KTH dataset show that the proposed approach outperforms existing deep models, and gives comparable results with the best related works.

Abstract. In this paper, we propose a novel approach for action classification in soccer videos using a recurrent neural network scheme. Thereby, we extract from each video action at each timestep a set of features which describe both the visual content (by the mean of a BoW approach) and the dominant motion (with a key point based approach). A Long Short-Term Memory-based Recurrent Neural Network is then trained to classify each video sequence considering the temporal evolution of the features for each timestep. Experimental results on the MICC-Soccer-Actions-4 database show that the proposed approach outperforms classification methods of related works (with a classification rate of 77 %), and that the combination of the two features (BoW and dominant motion) leads to a classification rate of 92 %.

No abstract

This paper presents a new robust method for inertial MEM (MicroElectroMechanical systems) 3D gesture recognition. The linear acceleration and the angular velocity, respectively provided by the accelerometer and the gyrometer, are sampled in time resulting in 6D values at each time step which are used as inputs for the gesture recognition system. We propose to build a system based on Bidirectional Long Short-Term Memory Recurrent Neural Networks (BLSTM-RNN) for gesture classification from raw MEM data. We also compare this system to a geometric approach using DTW (Dynamic Time Warping) and a statistical method based on HMM (Hidden Markov Model) from filtered and denoised MEM data. Experimental results on 22 individuals producing 14 gestures in the air show that the proposed approach outperforms classical classification methods with a classification mean rate of 95.57% and a standard deviation of 0.50 for 616 test gestures. Furthermore, these experiments underline that combining accelerometer and gyrometer information gives better results that using a single inertial description.

We present a novel strategy for unsupervised feature learning in image applications inspired by the Spike-Timing-Dependent-Plasticity (STDP) biological learning rule. We show equivalence between rank order coding Leaky-Integrate-and-Fire neurons and ReLU artificial neurons when applied to non-temporal data. We apply this to images using rank-order coding, which allows us to perform a full network simulation with a single feed-forward pass using GPU hardware. Next we introduce a binary STDP learning rule compatible with training on batches of images. Two mechanisms to stabilize the training are also presented : a Winner-Takes-All (WTA) framework which selects the most relevant patches to learn from along the spatial dimensions, and a simple feature-wise normalization as homeostatic process. This learning process allows us to train multi-layer architectures of convolutional sparse features. We apply our method to extract features from the MNIST, ETH80, CIFAR-10, and STL-10 datasets and show that these features are relevant for classification. We finally compare these results with several other state of the art unsupervised learning methods.