Faculty Publications

Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736

Publications by NITK Faculty

Browse

Subject: search.filters.subject.Long short-term memory

Search Results

Now showing 1 - 10 of 30
  • No Thumbnail Available
    Item
    Importance of Knee Angle and Trunk Lean in the Detection of an Abnormal Walking Pattern Using Machine Learning
    (Springer Science and Business Media Deutschland GmbH, 2023) Pandit, P.; Thummar, D.; Verma, K.; Gangadharan, K.V.; Das, B.; Kamat, Y.
    Human gait can be quantified using motion capture systems. Three-dimensional (3D) gait analysis is considered the gold standard for gait assessment. However, the process of three-dimensional analysis is cumbersome and time-consuming. It also requires complex software and a sophisticated environment. Hence, it is limited to a smaller section of the population. We, therefore, aim to develop a system that can predict abnormal walking patterns by analyzing trunk lean and knee angle information. A vision-based OpenPose algorithm was used to calculate individual trunk lean and knee angles. Web applications have been integrated with this algorithm so that any device can use it. A Miqus camera system of Qualisys 3D gait analysis system was used to validate the OpenPose algorithm. The validation method yielded an error of ± 9° in knee angle and ± 8° in trunk lean. The natural walking pattern of 100 healthy individuals was compared to simulated walking patterns in an unconstrained setting in order to develop a machine learning program. From the collected data, an RNN-based LSTM machine learning model was trained to distinguish between normal and abnormal walkings. LSTM-based models were able to distinguish between normal and abnormal gaits with an accuracy of 80%. This study shows that knee angle and trunk lean patterns collected during walking can be significant indicators of abnormal gait. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
  • No Thumbnail Available
    Item
    PhishDump: A multi-model ensemble based technique for the detection of phishing sites in mobile devices
    (Elsevier B.V., 2019) Rao, R.S.; Vaishnavi, T.; Pais, A.R.
    Phishing is a technique in which the attackers trick the online users to reveal the sensitive information by creating the phishing sites which look similar to that of legitimate sites. There exist many techniques to detect phishing sites in desktop computers. In recent years, the number of mobile users accessing the web has increased which lead to a rise in the number of attacks in mobile devices. Existing techniques designed for desktop computers may not be suitable for mobile devices due to their hardware limitations such as RAM, Screen size, low computational power etc. In this paper, we propose a mobile application named PhishDump to classify the legitimate and phishing websites in mobile devices. PhishDump is based on the multi-model ensemble of Long Short Term Memory (LSTM) and Support Vector Machine (SVM) classifier. As PhishDump focuses on the extraction of features from URL, it has several advantages over existing works such as fast computation, language independence and robust to accidental download of malwares. From the experimental analysis, we observed that our proposed multi-model ensemble outperformed traditional LSTM character and word-level models. PhishDump performed better than the existing baseline models with an accuracy of 97.30% on our dataset and 98.50% on the benchmark dataset. © 2019 Elsevier B.V.
  • No Thumbnail Available
    Item
    Efficient deep learning techniques for the detection of phishing websites
    (Springer, 2020) Somesha, M.; Pais, A.R.; Rao, R.S.; Rathour, V.S.
    Phishing is a fraudulent practice and a form of cyber-attack designed and executed with the sole purpose of gathering sensitive information by masquerading the genuine websites. Phishers fool users by replicating the original and genuine contents to reveal personal information such as security number, credit card number, password, etc. There are many anti-phishing techniques such as blacklist- or whitelist-, heuristic-feature- and visual-similarity-based methods proposed as of today. Modern browsers adapt to reduce the chances of users getting trapped into a vicious agenda, but still users fall as prey to phishers and end up revealing their secret information. In a previous work, the authors proposed a machine learning approach based on heuristic features for phishing website detection and achieved an accuracy of 99.5% using 18 features. In this paper, we have proposed novel phishing URL detection models using (a) Deep Neural Network (DNN), (b) Long Short-Term Memory (LSTM) and (c) Convolution Neural Network (CNN) using only 10 features of our earlier work. The proposed technique achieves an accuracy of 99.52% for DNN, 99.57% for LSTM and 99.43% for CNN. The proposed techniques utilize only one third-party service feature, thus making it more robust to failure and increases the speed of phishing detection. © 2020, Indian Academy of Sciences.
  • No Thumbnail Available
    Item
    Gearbox fault diagnosis based on Multi-Scale deep residual learning and stacked LSTM model
    (Elsevier B.V., 2021) Ravikumar, K.N.; Yadav, A.; Kumar, H.; Gangadharan, K.V.; Narasimhadhan, A.V.
    Fault diagnosis methods based on signal analysis techniques are widely used to diagnose faults in gear and bearing. This paper introduces a fault diagnosis model that includes a multi-scale deep residual learning with a stacked long short-term memory (MDRL-SLSTM) to address sequence data in a gearbox health prediction task in an internal combustion (IC) engine. In the MDRL-SLSTM network, CNN and residual learning is firstly utilized for local feature extraction and dimension reduction. The experiment is carried out on the gearbox of an IC engine setup, two datasets are used; one is from bearing and the other from 2nd driving gear of gearbox. To reduce the number of parameters, down-sampling is carried out on input data before giving to the architecture. The model achieved better diagnostic performance with vibration data of gearbox. Classification accuracy of 94.08% and 94.33% are attained on bearing datasets and 2nd driving gear of gearbox respectively. © 2021 Elsevier Ltd
  • No Thumbnail Available
    Item
    An effective feature extraction with deep neural network architecture for protein-secondary-structure prediction
    (Springer, 2021) Jayasimha, A.; Mudambi, R.; Pavan, P.; Lokaksha, B.M.; Bankapur, S.; Patil, N.
    With the increased importance of proteins in day-to-day life, it is imperative to know the protein functions. Deciphering protein structure elucidates protein functions. Experimental approaches for protein-structure analysis are expensive and time-consuming, and require high dexterity. Thus, finding a viable computational approach is vital. Due to the high complexity of predicting protein structure (tertiary structure) directly, research in this field aims at the protein-secondary-structure prediction which is directly related to its tertiary structure. This research aims at exploring a plethora of features, namely position-specific scoring matrices, hidden Markov model alignment matrices, and physicochemical properties, that carry rich information required to predict the secondary structure. Furthermore, it aims at exploring a suitable combination of the features which could capture diverse information about the protein secondary structure. Finally, a cascaded convolutional neural network and bidirectional long short-term memory architecture is fit on the models, and two evaluation metrics, namely, Q8 score and segment overlap score, are benchmarked on various datasets. Our proposed model trained on data of CB6133 dataset and tested on CB513 dataset beats the benchmark models by a minimum of 2.9%. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH Austria, part of Springer Nature.
  • No Thumbnail Available
    Item
    Human identification system using 3D skeleton-based gait features and LSTM model
    (Academic Press Inc., 2022) Rashmi, M.; Guddeti, R.M.R.
    Vision-based gait emerged as the preferred biometric in smart surveillance systems due to its unobtrusive nature. Recent advancements in low-cost depth sensors resulted in numerous 3D skeleton-based gait analysis techniques. For spatial–temporal analysis, existing state-of-the-art algorithms use frame-level information as the timestamp. This paper proposes gait event-level spatial–temporal features and LSTM-based deep learning model that treats each gait event as a timestamp to identify individuals from walking patterns observed in single and multi-view scenarios. On four publicly available datasets, the proposed system stands superior to state-of-the-art approaches utilizing a variety of conventional benchmark protocols. The proposed system achieved a recognition rate of greater than 99% in low-level ranks during the CMC test, making it suitable for practical applications. The statistical study of gait event-level features demonstrated retrieved features’ discriminating capacity in classification. Additionally, the ANOVA test performed on findings from K folds demonstrated the proposed system's significance in human identification. © 2021 Elsevier Inc.
  • No Thumbnail Available
    Item
    Time series forecasting of temperature and turbidity due to global warming in river Ganga at and around Varanasi, India
    (Springer Science and Business Media Deutschland GmbH, 2022) Das, N.; Sagar, A.; Bhattacharjee, R.; Agnihotri, A.K.; Ohri, A.; Gaur, S.
    The fluctuation in the river ecosystem network due to climate change-induced global warming affects aquatic organisms, water quality, and other ecological processes. Assessment of climate change-induced global warming impacts on regional hydrological processes is vital for effective water resource management and planning. The global warming effect on river water quality has been analyzed in this work. The river Ganga stretch near the Varanasi region has been chosen as the study area for this analysis. The air temperature has been predicted using the seasonal autoregressive integrated moving average (SARIMA) and the Prophet model. The Prophet model has shown better accuracy with a root mean square percent error (RMSPE) value of 3.2% compared to the SARIMA model, which has an RMPSE value of 7.54%. The river temperature, turbidity, and nighttime radiance values have been predicted for the years 2022 and 2025 using the long short-term memory (LSTM) algorithm. The anthropogenic effect on the river has been evaluated by using the nighttime radiance imageries. The predicted average river temperature shows an increment of 0.58 °C and 0.63 °C for the city and non-city river stretches, respectively, in 2025 compared to 2022. Similarly, the river turbidity shows an increment of 1.21 nephelometric turbidity units (NTU) and 1.17 NTU for the city and non-city stretch, respectively, in 2025 compared to 2022. For future predicted years, the nighttime radiance values for the region situated near the city river stretch show a significant rise compared to the region that lies nearby the non-city river stretch. © 2022, The Author(s), under exclusive licence to Springer Nature Switzerland AG.
  • No Thumbnail Available
    Item
    Semantic context driven language descriptions of videos using deep neural network
    (Springer Science and Business Media Deutschland GmbH, 2022) Naik, D.; Jaidhar, C.D.
    The massive addition of data to the internet in text, images, and videos made computer vision-based tasks challenging in the big data domain. Recent exploration of video data and progress in visual information captioning has been an arduous task in computer vision. Visual captioning is attributable to integrating visual information with natural language descriptions. This paper proposes an encoder-decoder framework with a 2D-Convolutional Neural Network (CNN) model and layered Long Short Term Memory (LSTM) as the encoder and an LSTM model integrated with an attention mechanism working as the decoder with a hybrid loss function. Visual feature vectors extracted from the video frames using a 2D-CNN model capture spatial features. Specifically, the visual feature vectors are fed into the layered LSTM to capture the temporal information. The attention mechanism enables the decoder to perceive and focus on relevant objects and correlate the visual context and language content for producing semantically correct captions. The visual features and GloVe word embeddings are input into the decoder to generate natural semantic descriptions for the videos. The performance of the proposed framework is evaluated on the video captioning benchmark dataset Microsoft Video Description (MSVD) using various well-known evaluation metrics. The experimental findings indicate that the suggested framework outperforms state-of-the-art techniques. Compared to the state-of-the-art research methods, the proposed model significantly increased all measures, B@1, B@2, B@3, B@4, METEOR, and CIDEr, with the score of 78.4, 64.8, 54.2, and 43.7, 32.3, and 70.7, respectively. The progression in all scores indicates a more excellent grasp of the context of the inputs, which results in more accurate caption prediction. © 2022, The Author(s).
  • No Thumbnail Available
    Item
    A novel Multi-Layer Attention Framework for visual description prediction using bidirectional LSTM
    (Springer Science and Business Media Deutschland GmbH, 2022) Naik, D.; Jaidhar, C.D.
    The massive influx of text, images, and videos to the internet has recently increased the challenge of computer vision-based tasks in big data. Integrating visual data with natural language to generate video explanations has been a challenge for decades. However, recent experiments on image/video captioning that employ Long-Short-Term-Memory (LSTM) have piqued the interest of researchers studying its possible application in video captioning. The proposed video captioning architecture combines the bidirectional multilayer LSTM (BiLSTM) encoder and unidirectional decoder. The innovative architecture also considers temporal relations when creating superior global video representations. In contrast to the majority of prior work, the most relevant features of a video are selected and utilized specifically for captioning purposes. Existing methods utilize a single-layer attention mechanism for linking visual input with phrase meaning. This approach employs LSTMs and a multilayer attention mechanism to extract characteristics from movies, construct links between multi-modal (words and visual material) representations, and generate sentences with rich semantic coherence. In addition, we evaluated the performance of the suggested system using a benchmark dataset for video captioning. The obtained results reveal superior performance relative to state-of-the-art works in METEOR and promising performance relative to the BLEU score. In terms of quantitative performance, the proposed approach outperforms most existing methodologies. © 2022, The Author(s).
  • No Thumbnail Available
    Item
    Airborne Kite Tether Force Estimation and Experimental Validation Using Analytical and Machine Learning Models for Coastal Regions
    (MDPI, 2022) Castelino, R.V.; Kashyap, Y.; Kosmopoulos, P.
    Wind power can significantly contribute to the transition from fossil fuels to renewable energies. Airborne Wind Energy (AWE) technology is one of the approaches to tapping the power of high-altitude wind. The main purpose of a ground-based kite power system is to estimate the tether force for autonomous operations. The tether force of a particular kite depends on the wind velocity and the kite’s orientation to the wind vector in the figure-eight trajectory. In this paper, we present an experimental measurement of the pulling force of an Airush Lithium 12 (Formula presented.) kite with a constant tether length of 24 m in a coastal region. We obtain the position and orientation data of the kite from the sensors mounted on the kite. The flight dynamics of the kite are studied using multiple field tests under steady and turbulent wind conditions. We propose a physical model (PM) using Artificial Neural Network (ANN) and Long Short-Term Memory (LSTM) deep neural network algorithms to estimate the tether force in the experimental validation. The performance study using the root mean square error (RMSE) method shows that the LSTM model performs better, with overall error values of 126 N and 168 N under steady and turbulent wind conditions. © 2022 by the authors.