K.N.V. SATYANARAYANA
@srkrec.edu.in
ASSISTANT PROFESSOR and Electronics and Communication Engineering College
SRKR Engineering College Bhimavaram.
K.N.V. Satyanarayana presently working as an assistant professor in Department of Electronics and Communication Engineering, S.R.K.R. Engineering College, Bhimavaram, A.P, India. He is currently pursuing PhD from Annamalai University. His current research interests include Image processing, Signal processing, Machine learning and Internet of things.
EDUCATION
Mtech.
RESEARCH INTERESTS
image Processing ,signal processing ,IoT ,Machine learning,
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Kalyanapu Jagadeeshwar, T. Sreenivasarao, Padmaja Pulicherla, K. N. V. Satyanarayana, K. Mohana Lakshmi, and Pala Mahesh Kumar
World Scientific Pub Co Pte Ltd
Automatic speech emotion recognition (ASER) from source speech signals is quite a challenging task since the recognition accuracy is highly dependent on extracted features of speech that are utilized for the classification of speech emotion. In addition, pre-processing and classification phases also play a key role in improving the accuracy of ASER system. Therefore, this paper proposes a deep learning convolutional neural network (DLCNN)-based ASER model, hereafter denoted with ASERNet. In addition, the speech denoising is employed with spectral subtraction (SS) and the extraction of deep features is done using integration of linear predictive coding (LPC) with Mel-frequency Cepstrum coefficients (MFCCs). Finally, DLCNN is employed to classify the emotion of speech from extracted deep features using LPC-MFCC. The simulation results demonstrate the superior performance of the proposed ASERNet model in terms of quality metrics such as accuracy, precision, recall, and F1-score, respectively, compared to state-of-the-art ASER approaches.
R. Saranya, M. Vijayaragavan, S. Ismail Kalilulah, K.N.V. Satyanarayana, M. Suresh, and S Srimathi
IEEE
Facial Emotion Recognition (FER) utilizing Deep Learning (DL) is indeed employed in the context of autonomous vehicle drivers. By investigating the facial expressions of drivers, the vehicle system is gain insights into the emotional state of drivers. FER is given to whole safety by identifying critical emotions like fear, anger, or surprise. When the system identifies these emotions, it responds accordingly by altering the driving performance or taking preventive measures for avoiding potential risks or aggressive maneuvers. The training of the DL algorithms namely Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) are done on the preprocessed data. These approaches learn to extract appropriate features in facial images and forecast the same emotions. This manuscript offers the design of Automated Facial Emotion Detection using Arithmetic Optimization Algorithm with Deep Convolutional Neural Network (AFED-AOADCNN) technique for Autonomous Vehicle Drivers. The purpose of the AFEDAOADCNN technique is to detect various kinds of facial emotions in autonomous vehicle drivers. In this introduced AFED-AOADCNN technique, DCNN method is applied for procedure of feature extraction. Next, the AOA is employed for optimum hyperparameter tuning of the DCNN method. At last, quantized neural network (QNN) approach is exploited for the identification and classification of different kinds of facial emotions. The investigational evaluation of the AFEDAOADCNN method is assessed on facial emotion dataset. The experimental validation stated the improved outputs of the AFED-AOADCNN method over recent approaches.
Satyanarayana Naga V. Kanuboyina, Shankar T, and Rama Raju Venkata Penmetsa
Informa UK Limited
K. N. V Satyanarayana, T. Shankar, G. Poojita, G Vinay, H. N. S. V. l Suvarna Amaranadh, and A. Gourisankar Babu
IEEE
Emotions are the requisites in our day-to-day life. Emotions are the psychophysiology states that are coupled with thoughts, feelings, behavioral responses, and a degree of satisfaction or dissatisfaction. There are various methods for achieving psychophysiology data from human beings, such as Electroencephalography (EEG), Electrocardiography (ECG), Photoplethysmogram (PPG), blood volume pulse (BVP). In this paper, the EEG de vices are considered for getting this data. Electroencephalography (EEG) is an electrophysiological monitoring method to note the electrical activity of the brain by the electrodes that are placed on the scalp. With the help of the deep dataset, the Support Vector Machine (SVM), which is a classifier is trained. The raw EEG data should be further processed to reduce the artifacts and features are selected to give the input to the SVM classifier. The outputs are in the form of valency and arousal values. The acquired results have an accuracy of 83% in the detection of emotions.
K. N. V Satyanarayana, V. Tejasri, Y. S. Naga. Srujitha, K. Nitya Sai Mounisha, Sai Tejasri Yerramsetti, and Gayathri Devi Darapu
IEEE
Emotions which can be commonly called to be as human feelings, are variable and numerous. They vary according to the situation or according to perception. Analyzing and classifying those emotions are very crucial in current situations. For example, for knowing the review of the product, the developer can use this emotion detection to see whether the client is satisfied with the product and can understand the likeliness of the product. Accordingly, he can vary it, and in health care for finding the depression in a person. So, this makes the classification of human feelings more vulnerable. Here initially, the data is being collected from the brain via EEG Signals, and fed into a mock dataset, and then these EEG Signal features can be extracted by using KNN Classifier to classify the data but To improve several parameters like time of execution and accuracy this seed data can be classified using the RNN(recurrent neural networks). For a small dataset, K nearest neighbour may work efficiently, but for large datasets and more classifications, a Recurrent neural network is more efficient. Here when a small seed dataset is being considered, It produces good accuracy and classification of the data. Computing using this process produces the best accuracy of 96.22% by the KNN classifier and Test accuracy of 85.71% by Recurrent Neural Networks.
K. N. V Satyanarayana, P. V Ramaraju, K. N. V Suresh Varma, and Y Ramalaxmanna
American Scientific Publishers
Mainly image processing is used for detection of objects with feasible number of constraints with different detection meth-odologies is defined used by camera-based detection. This method is used to find correspondence with respect to different objects. So that, in this paper, we propose Novel and simple method which is worked based on different region of interests present in video or image. This method mainly worked based on Histogram Oriented Gradients in image processing events. Our method also uses filtering approach with sequential data presentation to access interested data from image or video. Our experimental results mainly show effective visualization results with respect to different selection of regions.
K N. V. Satyanarayana, G Yaswanthini, P L. Kartheeka, and N Rajkumar
Science Publishing Corporation
Now-a-days road accidents are occurring frequently, due to rash driving of people. The most unfortunate thing is that by making small mistakes during driving, we lost our valuable future. If we observe, most of the accidents will occur at school zones, parks, hospitals, hill areas and highways. Even a police also can’t monitor all such kind of accidents. So in order to reduce the number of accidents and to control the vehicle speed the highway department has placed the signboards. But it is difficult to observe such kind of signboards and hence accidents will occur. This paper will provide a new way for controlling the speed of the vehicle without harming others. In this paper, we are using RFID module to limit vehicle speed. The RF transmitter will be placed at first and last of the restricted areas and RFID receiver should be placed inside the vehicle. The vehicle speed was obtained by speedometer which is available in vehicle. And that speed is compared and monitored by the controller. If the vehicle speed exceeds the limited speed, It automatically controls the speed of the vehicle according to that particular zone. Hence, automatically the speed reduced. If there is any emergency, a switch will be available in the vehicle. When the switch is ON, the speed is not controlled automatically. The vehicle which is switched ON, that vehicle number was stored in cloud. Here the main purpose of cloud is it loads the route map of the vehicle.