G R SURESH
@saveetha.com
Professor, ECE
Saveetha School of Engineering
Scopus Publications
Scholar Citations
Scholar h-index
Scholar i10-index
Scopus Publications
Gunduboyina Sairam Yadav, T. Sathish, and G. R. Suresh
AIP Publishing
C. Sasikumar, K. Malathi, and G. R. Suresh
AIP Publishing
C. Sasikumar and G. R. Suresh
AIP Publishing
P. Bindhu., G. R. Suresh, S. Rajalakshmi, C. Selvi, and S. Prakash
AIP Publishing
T. Rajesh Kumar, G. R. Suresh, and K. Kalaiselvi
World Scientific Pub Co Pte Ltd
Automatic Speech Recognition schemes are the alternating modes in which individuals interrelate with various mobile application applications. The user interactivity needs are a huge vocabulary identification system, high accuracy, energy-efficient solution and time. Although the automatic Speech Recognition system requires power budget and huge memory bandwidth, it is not applicable for many tiny forms of battery-controlled devices. Hence, an effective method is developed using the proposed Multi-kernel Perceptual linear predictive and Stochastic Biogeography-based whale optimization algorithm optimization to adapt non-audible mutter to regular speech. First, the input speech signal is initially given to the pre-processed module. Then, the features, such as spectral centroid, pitch chroma, Taylor amplitude modulation spectrogram (AMS), spectral skewness, and the developed Multi-kernel Perceptual linear predictive, are extracted to determine the appropriate features. After extracting features, the speech recognition is performed based on Deep Convolutional Neural Network, which is trained by the proposed Stochastic Biogeography whale optimization algorithm. The Stochastic Biogeography whale optimization algorithm combines the stochastic gradient descent method, whale optimization algorithm, and biogeography-based optimization. The developed model showed improved results with maximum accuracy of 0.985, minimal FPR of 0.001, maximal TPR of 1, respectively.
S. Jaanaa Rubavathy, G R Suresh, C. Senthilkumar, P. Shyamala Bharathi, and V. Amudha
IEEE
Respiratory rhythms are critical in a variety of medical emergencies. Respiratory rhythms of clinical relevance may be detected using a non-invasive respiratory analysis device established in this study. Light-weight wireless sensor nodes attached to the chest and belly of 150 healthy participants were used to gather data on controlled breathing. We next created our own datasets by infusing portions of quiet inhalation with annotated samples of different patterns. For each test datasets, with the one deep neural network has been used to locate the position of every occurrence and categorize it as corresponding to the one of the above-mentioned event kinds. For quiet inhalation, we got a mean F1 score of 93%, for central sleep apnea. Supervised learning may be used to interpret the data through sensing devices, such as chest and abdominal movement, to give a nonintrusive system to measure respiratory rate. Recognizing apneas when sleeping at night and measuring respiratory episodes in ventilated patients medical and surgical patients unit might benefit from this technology.
V. Amudha, S. Jaanaa Rubavathy, G R Suresh, C. Senthilkumar, and P. Shyamala Bharathi
IEEE
Wireless sensor networks have a substantial challenge in the form of excessive energy consumption brought on by the transmission of superfluous data (WSNs). The resolution of this issue results in an increase in the lifespan of any network and an improvement in the practicality of using the system for implementations. Expanding on this, a demand for electricity data gathering is becoming increasingly important for wireless sensor network applications that consist of low-powered sensing devices. In such situations, information grouping and modeling techniques that make use of symmetrical correlation inside the various sensors may be employed for the purpose of decreasing the amount of power that detectors need in order to continue collecting stored information. In this study, we propose a hybrid model called DTKFA that is premised on autoregressive integrated moving average, decision tree and Kalman filtering techniques. The aim of this process is to anticipate the data gathering requirements of sensor nodes in order to cut down on unneeded information transfer. Segmentation and information consolidation to every source node are applied in the WSNs in order to make appropriate sampling forecasts in an effective manner. This is primarily done in order to decrease the computational overheads that are generated by the forecasting model. The results of experimental tests, similarities, and job evaluation carried out in a variety of contexts demonstrate that now the accuracy rate of with us methodology could indeed keep growing Gaussian and probabilistic based models should give higher energy proficiency due to a reduction in the amount of data transmission. This is possible because our method reduces the number of packets that need to be sent.
P. Shyamala Bharathi, V. Amudha, S. Jaanaa Rubavathy, G R Suresh, and C. Senthilkumar
IEEE
In light of the fact that the global population as a whole is becoming older and the effects that this will have, it is essential to ensure that people of retirement age may continue to be active and keep their sense of autonomy for a longer period of time. The purpose of this Physical Fitness system (PFS) is to encourage strolls across the city as a means of encouraging lives that are more healthy and active. Body local are used to combine data acquired by various types of sensors, which is then transferred to a server to enable careers in making educated choices while arranging physical activities for the people they are caring for. In this paper, the Physical Fitness system is described, and a summary is given on a test design that was constructed in order to evaluate the effectiveness of the currently deployed version and the important structures that make up that application. According to the findings, the Physical Fitness device has a strong performance in order to adequately to rebound from connection failures, its ability to upload data, the amount of battery life it consumes, and the electromagnetic range of its personal area network.
T. Rajesh Kumar, G. N. Balaji, D. Vijendra Babu, Soubraylu Sivakumar, K. Kalaiselvi, and G. R. Suresh
Springer Nature Singapore
S. Saraswathi, G. R. Suresh, and Jeevaa Katiravan
Springer Science and Business Media LLC
Cross T. Asha Wise, G. R. Suresh, M. Palanivelen, and S. Saraswathi
World Scientific Pub Co Pte Lt
Mounting electronics circuits on a plastic flexible substrate are pertinent for biosensing applications due to their resilient nature, minimal processing conditions, lightweight and low cost. Organic Field-Effect Transistors (OFET)-based amplifier for flexible biosensors have been proposed in this paper. To design flexible biosensing circuits, Metal Oxide Semiconductor Field-Effect Transistor (MOSFET) with Polycyclic Hydrocarbon is a suitable choice. It is a big challenge to build an organic circuit using graphene electrode due to its poor performance of [Formula: see text]-type OFET, therefore it is advisable to use Pentacene as [Formula: see text]- and [Formula: see text]-type Organic semiconductors. Pentacene being one among the foremost totally investigated conjugated organic molecules with a high application potential because the hole mobility in OFETs goes up to 0.2[Formula: see text]cm2/(Vs), which exceeds that of amorphous silicon. In biosignal process, the first and most important step is to amplify the biosignal for further processing. Operational Transconductance Amplifier (OTA) plays an essential role in biological signal measuring instruments like EEG, ECG, EMG modules which measure the heart, muscle and brain activities. The OTA designed using this OFET is adaptable for flexible sensor circuits and also it derives the transconductance of 67 which is similar to silicon OTA. The amplifier designed here gives unit gain of 42[Formula: see text]dB with a frequency of 195[Formula: see text]Hz which is suitable for low-frequency biosignal processing applications.
T. Rajesh Kumar, G. R Suresh, S. Kanaga Subaraja, and C. Karthikeyan
Wiley
A. Selvarani and G. R. Suresh
IEEE
The tongue is a major part of the human body to taste, speak and swallow the food. Tongue portion is directly connected with our internal organ. If any problem in that it reflects the effect through the tongue. Tongue center portion is connected with stomach, pancreas. Side portions are connected with liver. Tongue tip is connected with the heart, etc. In this study, an efficient Decision Support System is used to diagnosis the diabetes based on Characterization of tongue images by analyzing tongue colour distribution and texture for diabetic patient and healthy person. Comparative analysis of tongue image of DM and Healthy Person by using Kernel Ensemble Classification (KEC) method. Also Show the lab values of Diabetes.
V. Kaviya and G.R. Suresh
IEEE
Heart attacks take the maximum toll of human life all over the world. Heart disease is reportedly killing approximately 17 million people in the world, and a similar scene is seen in india, where 3 million people die because of CVDs (cardio-vascular diseases), which include heart attack and stroke. Patients suffering from heart diseases need a earlier detection system to have a urgent treatment before it's too late. More over some existing methodology helps in detecting heart attack only if the patient is under doctor or expert observation and hence it is not possible for all the cases. To overcome the above disadvantage we have developed a system which will help to provide earlier detection of myocardial infarction, which avoid overdue of medical treatment. The proposed system is combined with a data collection and computational algorithm which can collects the multiple data continuously, according to the health risk conscious in each patient. In case of heart attack the system sends alert to person device and doctor or medical assistance through GSM. So, that medical assistant can reach the patient with the help of GPS location and provide early medical treatment.
K. Kalaiselvi, G.R. Suresh, and V. Ravi
Wiley
Wireless body area network (WBAN) plays an important role in patient health care. The performance of this WBAN system is affected by link failures due to the presence of malicious sensor nodes. Hence, the detection and mitigation of this link failure is important for improving the efficiency of the WBAN system. This paper proposes a methodology for link failure detection using weight metric approach. The performance of the proposed methodology is analyzed in terms of packet delivery ratio (PDR), link failure detection latency, and link failure detection rate.
P. Malathi, G. R. Suresh, M. Moorthi, and N. R. Shanker
Springer Science and Business Media LLC
K. Kalaiselvi, G. R. Suresh, and V. Ravi
Springer Science and Business Media LLC
T. Rajesh Kumar, G.R Sursh, and S. Kanaga Subaraja
IEEE
In recent years, majority of people are affected with low voice, due to the biological variations in the environment and hereditary infections occurred in throat or vocal track. These people communicate with others in a murmured low voice utterances. This paper introduces a novel approach for the processing of non-audible murmur (NAM), with respect to a Full Rank Gaussian Mixture Model (FR-GMM) using non-parallel training adaptation method. Non-audible murmur means, lightly uttered soft voice that is very difficult to hear. The non-audible murmur is extracted directly from the soft tissues behind the ear, utilizing a body conductive non-audible murmur microphone made of stethoscope sensor. In the proposed technique, a FR-GMM is prepared with the traditional strategy while allowing the reference speaker voices using nonparallel training adaptation method. The Investigational analysis outputs prove that the proposed approach improving the natural quality of speech by 6% in word precision compared to the traditional approaches.
A. Selvarani and G. R. Suresh
Springer Science and Business Media LLC
P. Malathi, G. R. Sureshw, and M. Moorthi
IEEE
Laryngectomees lose their voice box after surgery and adapt various methods to restore their voice, one of them being Electrolaryngeal speech. The Electrolarynx suffers from producing natural speech by generating mechanical form of speech with suppressed unvoiced features, device and environment noise. This paper tends to remove the echo noise, device noise and environmental noise thereby enhancing the Electrolaryngeal speech to be more intelligible by spectral mapping using Gaussian Mixture Model (GMM) and auditory masking. The low frequency noise is masked by the pre-emphasised speech signal by determining the absolute threshold of masking. The spectral mapping technique using GMM based voice conversion in association with the source-filter model improves the voice quality and prosody. The objective and subjective evaluation measures, depict the significant enhancement of electrolaryngeal speech compared to previous enhancement methods which removed only low frequency noise and failed to include voice quality.
T. Rajesh Kumar, G. R. Suresh, and S. Kanaga Suba Raja
American Scientific Publishers
R. Senthamizh Selvi, G. R Suresh, and S. Kanaga Suba Raja
American Scientific Publishers
R. Senthamizh Selvi, R. Kishore, G. R. Suresh, and S. Kanaga Suba Raja
IEEE
Piracy on the internet has been a long plaguing problem and it has made digital watermarking schemes to play an important role in handling such issues. In this paper, an adaptive watermarking algorithm HSA-EMD is used. Each audio signal is divided into frames and decomposed into intrinsic mode function (IMF). Each IMF has its own number of extrema and minima. The watermark is embedded into the extrema of last IMF. Hence the location of watermark cannot be detected. The proposed scheme is highly robust against various attacks. It gives improved performance than the existing schemes. Thus watermarked IMF is less sensitive to common signal processing attacks.
R. Preetha, R. Bhanumathi, and G.R. Suresh
Informa UK Limited
ABSTRACT Brain tumor is one of the leading causes of death making tumor detection very important and challenging in the medical field. This paper describes tumor detection in medical images using immune feature weighted least squares-support vector machine (IFWLS-SVM). The challenge in brain tumor detection in magnetic resonance (MR) images is the existence of non-linearity in real data. Least squares-support vector machine (LS-SVM) is a conventional algorithm that has been applied to diagnose the detection problems in MR images and non-linear distribution in brain tumors. LS-SVM solves a linear system for a training algorithm instead of using quadratic programming in SVM. In conventional LS-SVM, each sample feature taken has equal importance for classification results, which does not give accurate results in real applications. In addition, parameters of LS-SVM and their kernel function prominently affect the classification result. An IFWLS-SVM has been used to optimize the kernel and tune the parameters of LS-SVM in this paper. Theoretical analysis and experimental results showed that IFWLS-SVM has better performance than other conventional algorithms.