Feature Selection for Speech Emotion Recognition in Spanish and Basque: On the Use of Machine Learning to Improve Human-Computer Interaction

Study of emotions in human-computer interaction is a growing research area. This paper shows an attempt to select the most significant features for emotion recognition in spoken Basque and Spanish Languages using different methods for feature selection. RekEmozio database was used as the experimental data set. Several Machine Learning paradigms were used for the emotion classification task. Experiments were executed in three phases, using different sets of features as classification variables in each phase. Moreover, feature subset selection was applied at each phase in order to seek for the most relevant feature subset. The three phases approach was selected to check the validity of the proposed approach. Achieved results show that an instance-based learning algorithm using feature subset selection techniques based on evolutionary algorithms is the best Machine Learning paradigm in automatic emotion recognition, with all different feature sets, obtaining a mean of 80,05% emotion recognition rate in Basque and a 74,82% in Spanish. In order to check the goodness of the proposed process, a greedy searching approach (FSS-Forward) has been applied and a comparison between them is provided. Based on achieved results, a set of most relevant non-speaker dependent features is proposed for both languages and new perspectives are suggested.

Contributions to the efficient use of general purpose coprocessors: kernel density estimation as case study

142 p.

Efficient excitations of radially and azimuthally polarized Nd 3+:YAG ceramic microchip laser by use of subwavelength multilayer concentric gratings composed of Nb2O5/SiO2

Cylindrical vector beams were produced from laser diode end-pumped Nd:YAG ceramic microchip laser by use of two types of subwavelength multilayer gratings as the axisymmetric-polarization output couplers respectively. The grating mirrors are composed of high- and low-refractive-index (Nb2O5/SiO2) layers alternately while each layer is shaped into triangle and concentric corrugations. For radially polarized laser output, the beam power reached 610mW with a polarization extinction ratio ( PER) of 61: 1 and a slope efficiency of 68.2%; for azimuthally polarized laser output, the beam power reached 626mW with a PER of 58: 1 and a slope efficiency of 47.6%. In both cases, the laser beams had near-diffraction limited quality. Small differences of beam power, PER and slope efficiency between radially and azimuthally polarized laser outputs were not critical, and could be minimized by further optimized adjustment to laser cavity and the reflectances of respective grating mirrors. The results manifested, by use of the photonic crystal gratings mirrors and end-pumped microchip laser configuration, CVBs can be generated efficiently with high modal symmetry and polarization purity. (C) 2008 Optical Society of America.