Conference Papers
Permanent URI for this collectionhttps://idr.nitk.ac.in/handle/123456789/28506
Browse
11 results
Search Results
Item SVM based methods for arrhythmia classification in ECG(2010) Kohli, N.; Verma, N.K.; Roy, A.In this study, Support Vector Machine (SVM) based methods have been used to classify the electrocardiogram (ECG) arrhythmias. Among various existing SVM methods, three well-known and widely used algorithms one-against-one, one-against-all, and fuzzy decision function are used here to distinguish between the presence and absence of cardiac arrhythmia and classifying them into one of the arrhythmia groups. The various types of arrhythmias in the Cardiac Arrhythmias ECG database chosen from University of California at Irvine (UCI) to train SVM, include ischemic changes (coronary artery disease), old inferior myocardial infarction, sinus bradycardy, right bundle branch block, and others. The results obtained through implementation of all three methods are thus compared as per their accuracy rate in percentages and the performance of the SVM classifier using one-against-all (OAA) method was found to be better than other techniques. ECG arrhythmia data sets are of generally complex nature and SVM based one-against-all method is found to be of vital importance for classification based diagnosing diseases pertaining to abnormal heart beats. ©2010 IEEE.Item Urban land cover classification using hyperspectral data(International Society for Photogrammetry and Remote Sensing, 2014) Hegde, G.; Mohammed Ahamed, J.M.; Hebbar, R.; Raj, U.Urban land cover classification using remote sensing data is quite challenging due to spectrally and spatially complex urban features. The present study describes the potential use of hyperspectral data for urban land cover classification and its comparison with multispectral data. EO-1 Hyperion data of October 05, 2012 covering parts of Bengaluru city was analyzed for land cover classification. The hyperspectral data was initially corrected for atmospheric effects using MODTRAN based FLAASH module and Minimum Noise Fraction (MNF) transformation was applied to reduce data dimensionality. The threshold Eigen value of 1.76 in VNIR region and 1.68 in the SWIR region was used for selection of 145 stable bands. Advanced per pixel classifiers viz., Spectral Angle Mapper (SAM) and Support Vector Machine (SVM) were used for general urban land cover classification. Accuracy assessment of the classified data revealed that SVM was quite superior (82.4 per cent) for urban land cover classification as compared to SAM (67.1 per cent). Selecting training samples using end members significantly improved the classification accuracy by 20.1 per cent in SVM. The land cover classification using multispectral LISS-III data using SVM showed lower accuracy mainly due to limitation of spectral resolution. The study indicated the requirement of additional narrow bands for achieving reasonable classification accuracy of urban land cover. Future research is focused on generating hyperspectral library for different urban features.Item Characterization of aspirated and unaspirated sounds in speech(Institute of Electrical and Electronics Engineers Inc., 2017) Ramteke, P.B.; Sadanand, A.; Koolagudi, S.G.; Pai, V.In this work, consonant aspiration and unaspiration phenomena are studied. It is known that, pronunciation of aspiration and unaspiration is characterized by the 'puff of air' released at the place of constriction in the vocal tract which is known as burst. Here, the properties of vowel immediately after the burst are studied for characterization of the burst. Excitation source signal estimated from the speech linear prediction residual is used for the task. The signal characteristics such as glottal pulse, duration of open, closed & return phases, slope of open & return phases, duration of burst, ratio of highest and lowest energies of signal and voice onset time (VOT) are explored to characterize aspiration and unaspiration. TIMIT English speech corpus is used to test the proposed approach. Random forest (RF) and support vector machine (SVMs) are used as classifiers to test the effectiveness of the features used for the task. An accuracy of 99.93% and 94.03% is achieved respectively. From the results, it is observed that the proposed features are robust in classifying the aspirated and unaspirated consonants. © 2017 IEEE.Item Voice activity detection from the breathing pattern of the speaker(Institute of Electrical and Electronics Engineers Inc., 2018) Ramakrishnan, A.G.; Krishnan, G.; Srivathsan, S.In this paper, we propose a method to perform voice activity detection using only the breathing signal of a speaker. Human breathing and speech production go hand in hand. Normal respiration and respiration during speech have a different profile. The former is generally symmetric as compared to an asymmetric profile in the case of respiration during speech. Impedance pneumography provides a mechanism to capture chest expansions and compressions due to breathing. We have recorded the breathing signal along with the speech audio for 44 subjects while they were speaking and quiet. We have classified cycles of breathing into two classes, namely during speech and normal, using the cycle-synchronous discrete cosine transform coefficients of the breathing signal with different classifiers. The best accuracy of 96.4% is obtained using the k-nearest neighbor classifier. From the classified breathing cycles, we determine the intervals when a subject is quiet and when he is speaking. We use the corresponding timeframes on the simultaneously recorded audio and achieve a good accuracy in voice activity detection. Compared to the earlier reported time resolution of 30 sec, we obtain a decision for every breathing cycle, which works out to an average resolution of about 3 sec. © 2017 IEEE.Item Detection of pathological condition of heart using texture complexity of the heart sound signals in kernel space(Institute of Electrical and Electronics Engineers Inc., 2018) Mondal, A.; Palaniappan, R.The heart sound signal is produced due to the mechanical activity of cardiovascular system and hence, it carries relevant information regarding the pathological and non-pathological condition of heart. Physicians listen to heart sounds using a stethoscope device and make interpretation regarding the diseases if any. Hence, its performance depends on the medical practitioners knowledge and experience. In this study, a very simple alternative approach of auscultation technique is introduced for early detection of cardiac abnormality based on a combined framework of kernel space decomposition and textural analysis. The experimental results are validated by physician's diagnosis and statistical analysis. The proposed diagnostic tool can be used in combination with audio clue of heart sound signal to train inexperienced physicians and as an assistive device for experienced doctors. © 2018 IEEE.Item Wearable sensor-based human fall detection wireless system(Springer Verlag service@springer.de, 2019) Kumar, V.S.; Acharya, K.G.; Bairampalli, B.; Thyagarajan, T.; Chaturvedi, A.Background/Objectives: Human fall detection is a critical challenge in the healthcare domain since the late medical salvage will even lead to death situations, therefore it requires timely rescue. This research work proposes a system which uses a wearable device that senses human fall and wirelessly raises alerts. Methods/statistical analysis: The detection system consists of the sensor system which contains both accelerometer and gyroscope sensors. The proper orientation of the subject is provided by the Madgwick filter. Six volunteers were engaged to perform the falling and non-falling events. The system is validated and checked by four algorithms: threshold based, support vector machine (SVM), K-nearest neighbor, and dynamic time wrapping, and thus, the accuracy was calculated. Findings: From the results obtained, the SVM has given an accuracy of 93%. Conclusions: When a fall is being detected, an additional feature to check whether the person is in critical state and is lying down for more than a particular time is incorporated and a critical alert is sent to the caretaker’s mobile. © Springer Nature Singapore Pte Ltd. 2019.Item Tea leaf disease prediction using texture-based image processing(Springer, 2020) Srivastava, A.R.; Venkatesan, M.Nowadays, Tea is commonly used in India as well as in all over the world. Tea is produced in many states of India, i.e., Assam, West Bengal, Tamil Nadu, Karnataka, and so on. But, production of tea is heavily affected by various diseases and pests. There are various kinds of diseases in tea leaves and various pests that can damage the tea crop and affect the tea production. Tea crop damage is reduced by recognizing the tea leaf diseases in an early stage. After detection of the kind of tea leaf diseases, suitable preventive method can be used to reduce the tea crop damage. For the detection of tea leaves diseases, there are different classification methods. Some classification techniques are random forest classifier, k-nearest neighbor classifier, support vector machine classifier, neural network, etc. After training the dataset with classifier, the image of tea leaf is given as an input, the best possible match is found by the classifier system, and diseases are recognized by the classifier system. This project is going to use various classification techniques to recognize and predict the tea leaves disease which helps us to improve the tea production of India. © Springer Nature Singapore Pte Ltd 2020.Item Identification of Palatal Fricative Fronting Using Shannon Entropy of Spectrogram(Springer Science and Business Media Deutschland GmbH, 2020) Ramteke, P.B.; Supanekar, S.; Aithal, V.; Koolagudi, S.G.In this paper, an attempt has been made to identify palatal fricative fronting in children speech, where postalveolar /sh/ is mispronounced as dental /s/. In children’s speech, the concentration of energy (darkest part) of spectrogram for /s/ ranges 4000 Hz to 8000 Hz, whereas it ranges 3000 Hz 8000 Hz for /sh/. Gammatonegram follows the frequency subbands of the ear (wider for higher frequencies). Various spectral properties such as spectral centroid, spectral crest factor, spectral decrease, spectral flatness, spectral flux, spectral kurtosis, spectral spread, spectral skewness, spectral slope and Shannon entropy of the spectrogram (interval of 2000 Hz), extracted from the Gammatonegram are proposed for the characterization of /sh/ and /s/. The dataset recorded from 60 native Kannada speaking children of age between 3 1/2 to 6 1/2 years is considered for the analysis from NITK Kids’ Speech Corpus. Support vector machine (SVMs) is considered for the classification. Various combinations of the proposed features are considered for the evaluation, along with the MFCCs(39) and LPCCs(39). Combination of MFCCs(39), LPCCs(39) and Entropy(4) is observed to achieve highest mispronunciation identification performance of 83.2983%. © 2020, Springer Nature Switzerland AG.Item Prevention of webshell attack using machine learning techniques(Grenze Scientific Society, 2021) Satish, Y.C.; Naik, P.M.; Rudra, B.Webshell is a web vulnerability and a security threat to any user or a server that can be accessed by attackers to control our system. And also, they may use our system as a command control device to attack other systems. It is difficult to monitor and identify such threats because attackers always tried to attack in different methods and new technologies. However, we can detect the webshell with Machine Learning Techniques with better accuracy; all we need is more number of samples. With this project, we presented a PHP based webshell detecting model. We used different ML algorithms: Logistic Regression(LR), Random Forest(RF), Support Vector Machine(SVM) and K-Nearest Neighbour(KNN). Addition to this PHP file's standard statistical features, we also added an opcode sequence from the PHP files, consisting of the TF-IDF Vector and the Hash Vector. Depending upon these features, we trained with different machine learning models(SVM, RF, LR, KNN). In these models, we got better results with Random Forest having an accuracy of 96.45\% with a false-positive rate of 3.5\%, which is good results compared to several popular detection techniques. © Grenze Scientific Society, 2021.Item A Machine Learning Approach for Daily Temperature Prediction Using Big Data(Springer Science and Business Media Deutschland GmbH, 2022) Divakarla, U.; Chandrasekaran, K.; Hemant Kumar Reddy, K.H.K.; Reddy, R.V.; Rao, M.Due to global warming, weather forecasting becomes complex problem which is affected by a lot of factors like temperature, wind speed, humidity, year, month, day, etc. weather prediction depends on historical data and computational power to analyze. Weather prediction helps us in many ways like in astronomy, agriculture, predicting tsunamis, drought, etc. this helps us to be prepared in advance for any kinds disasters. With rapid development in computational power of high end machines and availability of enormous data weather prediction becomes more and more popular. But handling such huge data becomes an issue for real time prediction. In this paper, we introduced the machine learning-based prediction approach in Hadoop clusters. The extensive use of map-reduce function helps us distribute the big data into different clusters as it is designed to scale up from single servers to thousands of machines, each offering local computation and storage. An ensemble distributed machine learning algorithms are employed to predict the daily temperature. The experimental results of proposed model outperform than the techniques available in literature. © 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.