A Novel Classification Method with Cubic Spline Interpolation

Classification is the last, and usually most time-consuming step in recognition. Most recently proposed classification algorithms have adopted machine learning (ML) as main approach, regardless of time consumption. This study proposes a statistical feature cubic spline interpolation (FC-CSI) algorithm to classify emotions speech using curve fitting technique. FC-CSI utilized emotion recognition system (SERS). The idea sketch (CSI) for each audio file dataset mean interpolations (MCSIs) representing dataset. CSI generated by connecting features extracted from extraction phase. MCSI 70% files Points on are considered new features. To according emotion, Euclidian distance (ED) found between all MCSIs Each classified nearest it. three datasets used this work Ryerson Audio-Visual Database Emotional Speech Song (RAVDESS), Berlin (Emo-DB), Surrey Expressed Emotion (SAVEE). shows fast high accuracy results. accuracy, i.e., proportion samples assigned correct class, without selection (FS), was 69.08%, 92.52%, 89.1% with RAVDESS, Emo-DB, SAVEE, respectively. results method were compared those designed neural network called SER-NN. Comparisons made FS. outperformed SER-NN Emo-DB underperformed an FS algorithm. It noticed experiments that operated faster than same utilizing