Facial expressions are universal and independent of race, culture, ethnicity, nationality, gender, age, religion, or any other demographic variable. These facts are the main reason for automatic facial expression reco...
详细信息
ISBN:
(纸本)9781509013531
Facial expressions are universal and independent of race, culture, ethnicity, nationality, gender, age, religion, or any other demographic variable. These facts are the main reason for automatic facial expression recognition being one of the hot topics of many research efforts and being useful in so many commercial and scientific fields. The most well-known and probably the most used anatomically based method of defining facial activity is Facial Action Coding System (FACS). In this paper, we propose a Facial Action Unit recognition algorithm using graph-based feature selection in unsupervised and supervised setting. The proposed algorithm is based on a state of the art algorithm for facial key points detection supervised gradient descent method, the classification is carried out using the well know Support Vector Machines classifier. Built this way, the algorithm works on still images where the human expressions are expected to be in their apex phase. Using leave one person out evaluation methodology we achieve average accuracy of 90.1% for unsupervised and 92.7% for supervised feature selection on 12 Action Units.
A novel recurrent wavelet-based Elman neural network (RWENN) control system is proposed in this study to control the mover position of a multi-axis motion control stage using linear ultrasonic motors (LUSMs) for the t...
详细信息
A novel recurrent wavelet-based Elman neural network (RWENN) control system is proposed in this study to control the mover position of a multi-axis motion control stage using linear ultrasonic motors (LUSMs) for the tracking of various contours. First, the structure and operating principles of the LUSMs are introduced briefly. Since the dynamic characteristics and motor parameters of the LUSMs are non-linear and time varying, the RWENN is proposed to control the mover of the X-Y-theta motion control stage to track various contours precisely using a direct decentralised control strategy. In the proposed RWENN, each hidden neuron employs a different wavelet function as an activation function. Moreover, the recurrent connective weights are added in the RWENN. Therefore compared with the conventional Elman neural network (ENN), both the precision and time of convergence are improved. Furthermore, the on-line learning algorithm based on the supervised gradient descent method and the convergence analysis of the tracking error using a discrete-type Lyapunov function of the RWENN are developed. Finally, some experimental results of various contours tracking show that the tracking performance of the RWENN is significantly improved compared with the ENN.
A Hermite polynomial-based recurrent neural network (HPBRNN) is proposed to control the rotor position on the axial direction of a thrust active magnetic bearing (TAMB) system for the tracking of various reference tra...
详细信息
A Hermite polynomial-based recurrent neural network (HPBRNN) is proposed to control the rotor position on the axial direction of a thrust active magnetic bearing (TAMB) system for the tracking of various reference trajectories in this study. First, the operating principles and dynamic model of the TAMB system using the non-linear electromagnetic force model is derived. Then, the HPBRNN is developed for the TAMB system with enhanced control performance and robustness. In the proposed HPBRNN, each hidden neuron employs a different orthonormal Hermite polynomial basis function (OHPBF) as an activation function. Therefore the learning ability of the HPBRNN is effective with high convergence precision and fast convergence time. Moreover, the connective weights of the HPBRNN using the supervised gradient descent method are updated online and the convergence analysis of the tracking error using the discrete-type Lyapunov function is provided. Finally, some experimental results of various reference trajectories tracking show that the control performance of the HPBRNN is significantly improved compared to the conventional proportional-integral-derivative and recurrent neural network controllers and demonstrate the validity of the proposed HPBRNN for practical applications.
暂无评论