Although deep neural networks (DNNs) have led to many remarkable results in cognitive tasks, they are still far from catching up with human-level cognition in antinoise capability. New research indicates how brittle and susceptible current models are to small variations in data distribution. In this letter, we study the stochasticity-resistance character of biological neurons by simulating the input-output response process of a leaky integrate-and-fire (LIF) neuron model and proposed a novel activation function, rand softplus, (RSP) to model the response process. In RSP, a scale factor η is employed to mimic the stochasticity-adaptability of biological neurons, thereby enabling the antinoise capability of a DNN to be improved by the novel activation function. We validated the performance of RSP with a 19-layer residual network (ResNet) and a 19-layer visual geometry group (VGG) on facial expression recognition data sets and compared it with other popular activation functions, such as rectified linear unit...

Address event representation (AER) vision sensors process visual information by monitoring changes in light intensity and generating event streams. In this paper, a threshold mechanism based on the statistical results of the peak membrane potentials, is proposed to improve the accuracy and efficiency of the motion symbol detection (MSD) method based on the leaky integrate-and-fire (LIF) neuron model and a peak spiking monitoring unit (PSMU). And a multi-kernel learning algorithm based on the tempotron rule, namely MK-tempotron, is proposed to improve the antinoise performance of the classifier. In MK-tempotron, different kernels are applied to calculate the post-synaptic membrane potentials (PSPs) according to different input spiking patterns, to counteract the impact of noise on the spiking activities, so that the synaptic weights can be changed in the direction of correct firing. We verified the effectiveness of the threshold mechanism and multi-kernel learning on a variety of data sets. The experimental results show that among the several recent algorithms based on temporal encoding and SNN classifier, which have advantages in power consumption and network latency, our method achieved the best accuracy on MNIST-DVS (100ms) and N-CARS, and it also achieved competitive results on N-MNIST, CIFAR10-DVS, Posture-DVS, Poker-DVS and MNIST-DVS (200ms/500ms/2000ms).

To solve the problems of absence of expression-correlation in single-image-based expression recognition, and limitation of one classifier, a novel facial expression recognition method based on reverse co-saliency region features is proposed in this paper. Firstly, seven different classifications of expression images of a same person are used to extract facial changed regions of different expressions by a reverse co-saliency features method, named Reverse Co-Salient Regions (RCSR). Secondly, the extracted salient regions are described as texture and shape features LBP and HOG. Finally, based on RCSR features, multi-classifiers decision mechanism is used for expressions recognition. The following experimental results show that the recognition accuracy increases obviously compared with 3 different kinds of single-image-based expression recognition methods.