Yazar "Muthusamy, Hariharan" seçeneğine göre listele
Listeleniyor 1 - 7 / 7
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Bispectral features and mean shift clustering for stress and emotion recognition from natural speech(Pergamon-Elsevier Science Ltd, 2017) C. K., Yogesh; Muthusamy, Hariharan; Rajamanickam, Yuvaraj; Ngadiran, Ruzelita; Adom, A. H.; Polat, KemalA new set of features and feature enhancement techniques are proposed to recognize emotion and stress from speech signal. The speech waveforms and the glottal waveforms (derived from the recorded emotional/stress speech waveforms) were processed by using third order statistics called bispectrum and 28 (14 from speech waveforms and 14 from glottal waveforms) bispectral based features. In this work, mean shift clustering was used to enhance the discrimination ability of the extracted Bispectral Features (BSFs). Four classifiers were used to distinguish different emotional and stressed states. The performance of the proposed method is tested with three databases. Different experiments were conducted and recognition rates were achieved in the range between 93.44% and 100% for Berlin emotional speech database (BES), between 73.81% and 97.23% for Surrey audio-visual expressed emotion database (SAVEE), between 93.8% and 100% for speech under simulated and actual stress simulated domain (SUSAS) (recognition of multi-style speech under stress-neutral, loud, lombard and anger) and 100% for SUSAS actual domain (recognition of three different levels of stress. high, medium and low). The obtained results indicate that the proposed bispectral based features and mean shift clustering provide promising results to recognize emotion and stress from speech signal. (C) 2017 Elsevier Ltd. All rights reserved.Öğe Feature extraetion for biometrie reeognition with photoplethysmography signals(2013) Kavsaoğlu, Ahmet Reşit; Polat, Kemal; Bozkurt, Mehmet Recep; Muthusamy, HariharanPhotoplethysmography (PPG) signals stand out due to features such as readily accessible, high reliability and confidentiality, the ease of use etc. among bio-signals. The feasibility studies carried out on the PPG signals demonstrated that PPG signals contained important features for human recognition and were the availability of biometric identification systems. In this study, twenty new features were extracted from PPG signal as a preliminary study intended to biometric recognition. PPG signals with 10 seconds were recorded from five healthy people using SDPPG (second derivative PPG) data acquisition card. To remove the noise from received raw PPG signals, a FIR low pass filtering with 200 points and 10 Hz cut-off frequency was designed. These twenty new features were obtained from filtered PPG signal and its second derivative. PPG signal with 10 seconds contains eight periods and twenty characteristic features in each person must not change within an individual over a period. This feature symbolizes the consistency in the identification of a person. To test the performance of biometrie recognition system, the k-NN (k-nearest neighbor) classifier was used and achieved 95% of recognition success rate using lO-fold cross validation with twenty new features. The obtained results showed that the developed biometric recognition system based on PPG signal were very promising. © 2013 IEEE.Öğe Guest editorial: new trends in data pre-processing methods for signal and image classification(Springer, 2017) Polat, Kemal; Muthusamy, Hariharan; Acharya, Rajendra; Guo, YanhuiA special issue of the Neural Computing and Applications (NCAA) is dedicated to ‘‘New trends in data pre-processing methods for signal and image classification.’’Data pre-processing is crucial for effective data mining. Low-quality data usually produce inaccurate and unpredictable outcomes. Today’s real-world data are greatly vulnerable to noise and getting lost due to either large data size or the sources of origin. Real-world data are often inconsistent and incomplete, and are possible to have several errors. These poor-quality data will result in poorquality mining outcomes. Data pre-processing enhances the data standard and subsequently aids to refine the value of data mining outcomes. Data pre-processing performs certain processing on raw original data to prepare it for further processing or analysis. In short, data pre-processing prepares original raw data for further processing. Data preprocessing converts the data into a form acceptable easily for further processing by the user.Öğe An improved elephant herding optimization using sine–cosine mechanism and opposition based learning for global optimization problems(Elsevier, 2021) Muthusamy, Hariharan; Ravindran, Sindhu; Yaacob, Sazali; Polat, KemalAn improved elephant herding optimization (EHOI) is proposed for continuous function optimization, financial stress prediction problem and two engineering optimization problems in this work. Elephant Herding Optimization (EHO) is a swarm-based algorithm and was inspired by the social behaviour of elephant clans. In the literature, EHO has received great attention from researchers due to its global optimization capability and ease of implementation. However, it has few limitations like random replacing of worst individual and lack of exploitation, which leads to slow convergence. In this work, EHO was enhanced with the help of the position updating mechanism of sine–cosine algorithm (SCA) and opposition-based learning (OBL). The separating operator in original EHO was replaced by the sine–cosine mechanism and followed by opposition-based learning was introduced to increase the performance of EHO. The proposed EHOI was compared with eight well-known meta-heuristic optimization algorithms (MAs) by using 23 classical benchmark functions, 10 modern CEC2019 benchmark test functions and two engineering optimization problems. From the results, it was observed that the proposed EHOI outperformed most of the selected MAs in terms of solution quality. A kernel extreme learning machine (KELM) model was optimized by improved EHO and applied to handle financial stress prediction. The efficiency of the proposed EHOI_KELM model was tested on two popular financial datasets and compared with popular classifiers, EHO_KELM and SCA_KELM models. The results demonstrate that the proposed EHOI_KELM model shows excellent performance than the popular classifiers, EHO_KELM & SCA_KELM models and it can also serve as an effective tool for financial prediction.Öğe Improved emotion recognition using Gaussian mixture model and extreme learning machine in speech and glottal signals(Hindawi Ltd, 2015) Muthusamy, Hariharan; Polat, Kemal; Yaacob, SazaliRecently, researchers have paid escalating attention to studying the emotional state of an individual from his/her speech signals as the speech signal is the fastest and the most natural method of communication between individuals. In this work, new feature enhancement using Gaussian mixture model (GMM) was proposed to enhance the discriminatory power of the features extracted from speech and glottal signals. Three different emotional speech databases were utilized to gauge the proposed methods. Extreme learning machine (ELM) and k-nearest neighbor (kNN) classifier were employed to classify the different types of emotions. Several experiments were conducted and results show that the proposed methods significantly improved the speech emotion recognition performance compared to research works published in the literature.Öğe Particle swarm optimization based feature enhancement and feature selection for improved emotion recognition in speech and glottal signals(Public Library Science, 2015) Muthusamy, Hariharan; Polat, Kemal; Yaacob, SazaliIn the recent years, many research works have been published using speech related features for speech emotion recognition, however, recent studies show that there is a strong correlation between emotional states and glottal features. In this work, Mel-frequency cepstralcoefficients (MFCCs), linear predictive cepstral coefficients (LPCCs), perceptual linear predictive (PLP) features, gammatone filter outputs, timbral texture features, stationary wavelet transform based timbral texture features and relative wavelet packet energy and entropy features were extracted from the emotional speech (ES) signals and its glottal waveforms(GW). Particle swarm optimization based clustering (PSOC) and wrapper based particle swarm optimization (WPSO) were proposed to enhance the discerning ability of the features and to select the discriminating features respectively. Three different emotional speech databases were utilized to gauge the proposed method. Extreme learning machine (ELM) was employed to classify the different types of emotions. Different experiments were conducted and the results show that the proposed method significantly improves the speech emotion recognition performance compared to previous works published in the literature.Öğe Preface(Elsevier Science Bv, 2015) Polat, Kemal; Asyalı, Musa Hakan; Muthusamy, Hariharan; Acharya, U. Rajendra; Koeppen, Mario
