Faculty Publications
Permanent URI for this communityhttps://idr.nitk.ac.in/handle/123456789/18736
Publications by NITK Faculty
Browse
16 results
Search Results
Item Deep Learning for COVID-19(Springer Science and Business Media Deutschland GmbH, 2022) Bs, B.S.; Manoj Kumar, M.V.; Thomas, L.; Ajay Kumar, M.A.; Wu, D.; Annappa, B.; Hebbar, A.; Vishnu Srinivasa Murthy, Y.V.S.Ever since the outbreak in Wuhan, China, a variant of Coronavirus named “COVID 19” has taken human lives in millions all around the world. The detection of the infection is quite tedious since it takes 3–14 days for the symptoms to surface in patients. Early detection of the infection and prohibiting it would limit the spread to only to Local Transmission. Deep learning techniques can be used to gain insights on the early detection of infection on the medical image data such as Computed Tomography (CT images), Magnetic resonance Imaging (MRI images), and X-Ray images collected from the infected patients provided by the Medical institution or from the publicly available databases. The same techniques can be applied to do the analysis of infection rates and do predictions for the coming days. A wide range of open-source pre-trained models that are trained for general classification or segmentation is available for the proposed study. Using these models with the concept of transfer learning, obtained resultant models when applied to the medical image datasets would draw much more insights into the COVID-19 detection and prediction process. Innumerable works have been done by researchers all over the world on the publicly available COVID-19 datasets and were successful in deriving good results. Visualizing the results and presenting the summarized data of prediction in a cleaner, unambiguous way to the doctors would also facilitate the early detection and prevention of COVID-19 Infection. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.Item Multiple Choice Question Answering Using Attention Based Ranking and Transfer Learning(Institute of Electrical and Electronics Engineers Inc., 2022) Kadam, N.; Anand Kumar, A.M.The multiple choice question answering is still considered as an challenging task in Natural Language Processing. In this paper, we have tried to solve the problem of answering multiple choice questions where supporting documents corresponding to each question are not explicitly provided. Context retrieval is the strategy, which focuses on both reasoning and retrieving better supporting contexts. We present a improvised version of attention based deep neural network that eventually learns to order documents according to their relevance in relation to a given topic, all while achieving the goal of predicting the correct response. The top documents retrieved are considered more relevant context for given question answer pair. To achieve more accurate results transformer based pre-trained models are used in the implementation. We have used the concept of transfer learning which is related to learning and adapting knowledge by fine tuning model on other datasets. The reasoning challenge dataset by Allen institute is used to test the approach and SQuAD 2.0 and RACE datasets are used to fine tune the transformer based models. The accuracy of proposed model on ARC easy dataset is 89.51% and on ARC challenge dataset is 62.53%. © 2022 IEEE.Item An Approach for Waste Classification Using Data Augmentation and Transfer Learning Models(Springer Science and Business Media Deutschland GmbH, 2023) Kumsetty, N.V.; Bhat Nekkare, A.B.; Kamath S․, S.; Anand Kumar, M.Waste segregation has become a daunting problem in the twenty-first century, as careless waste disposal manifests significant ecological and health concerns. Existing approaches to waste disposal primarily rely on incineration or land filling, neither of which are sustainable. Hence, responsible recycling and then adequate disposal is the optimal solution promoting both environment-friendly practices and reuse. In this paper, a computer vision-based approach for automated waste classification across multiple classes of waste products is proposed. We focus on improving the quality of existing datasets using data augmentation and image processing techniques. We also experiment with transfer learning based models such as ResNet and VGG for fast and accurate classification. The models were trained, validated, and tested on the benchmark TrashNet and TACO datasets. During experimental evaluation, the proposed model achieved 93.13% accuracy on TrashNet and outperformed state-of-the-art models by a margin of 16% on TACO. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.Item An Efficient Deep Transfer Learning Approach for Classification of Skin Cancer Images(Springer Science and Business Media Deutschland GmbH, 2023) Naik, P.P.; Annappa, B.; Dodia, S.Prolonged exposure to the sun for an extended period can likely cause skin cancer, which is an abnormal proliferation of skin cells. The early detection of this illness necessitates the classification of der-matoscopic images, making it an enticing study problem. Deep learning is playing a crucial role in efficient dermoscopic analysis. Modified version of MobileNetV2 is proposed for the classification of skin cancer images in seven classes. The proposed deep learning model employs transfer learning and various data augmentation techniques to more accurately classify skin lesions compared to existing models. To improve the per¬formance of the classifier, data augmentation techniques are performed on “HAM10000" (Human Against Machine) dataset to classify seven dif¬ferent kinds of skin cancer. The proposed model obtained a training accuracy of 96.56% and testing accuracy of 93.11%. Also, it has a lower number of parameters in comparison to existing methods. The aim of the study is to aid dermatologists in the clinic to make more accurate diagnoses of skin lesions and in the early detection of skin cancer. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.Item A Comparative Study on End-to-End Learning for Self-Driving Cars(Springer Science and Business Media Deutschland GmbH, 2024) Kumar, S.; Pir, M.A.; Rajan, J.; Talawar, B.Autonomous vehicle technology has advanced in recent years. The self-driving car is one of the most attractive research fields, and automakers are fast focusing on it. There have been a number of attempts made in this field, such as lane recognition, the detection of objects on roadways, and the reconstruction of three-dimensional models; however, the focus of our study is on models that directly transform the camera input images into steering angles. In this paper, we performed a comparative study of some of the popular end-to-end CNN models pertaining to autonomous vehicles. We used four different data sets for model training and validation. Only one of the data sets was gathered from the real world; the other three were created using software simulations. For evaluating the performance of different models, we used the mean squared error (MSE) metric. It was interesting to see that certain models fared better than others when applied to diverse data sets. When considering real-world datasets, both pre-trained VGG-16 and pre-trained VGG-19 using transfer learning exhibit comparable performance, achieving an MSE value of 21.4 which is better than all other considered models. However, in the case of simulated datasets, pre-trained VGG-19 outperforms the majority of the other models. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.Item A Deep Learning Approach to Enhance Semantic Segmentation of Bacteria and Pus Cells from Microscopic Urine Smear Images Using Synthetic Data(Springer Science and Business Media Deutschland GmbH, 2024) Kanabur, V.R.; Vijayasenan, D.; Sumam David, S.; Govindan, S.Urine smear analysis aids in preliminary diagnosis of Urinary Tract Infection. But it is time-consuming and requires a lot of medical expertise. Automating the process using machine learning can save time and effort. However obtaining a large medical dataset is difficult due to data privacy concerns and medical expertise requirements. In this study, we propose a method to synthesize a large dataset of gram-stained microscopic images containing pus cells and bacteria. We train a machine learning model to achieve semantic segmentation of bacteria and pus cells using this dataset. Later we use it to perform transfer learning on a relatively small dataset of gram stained urine microscopic images. Our approach improved the F1-score from 50% to 63% for bacteria segmentation and from 77% to 83% for pus cell segmentation. This method has the potential to improve the turn-around time and the quality of preliminary diagnosis of Urinary Tract Infection. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2024.Item Enhancing Paediatric Healthcare: Deep Learning-Based Pneumonia Diagnosis from Children's Chest X-rays(Association for Computing Machinery, 2024) Patidar, M.; Pandey, G.; Koolagudi, S.G.; Karanth, K.S.; Chandra, V.Pneumonia is a severe disease in children and adults caused by lung infection. It is also the major cause of death in young children. Early diagnosis of pneumonia is essential as it can be life-threatening if not treated at the right time. In this paper, pneumonia detection in children using chest X-ray images has been done. The dataset considered for this work is the Kermany pneumonia chest X-ray dataset and a newly collected high-resolution dataset by us of children's Chest X-rays from Father Muller Hospital Mangalore, Karnataka. The dataset consists of chest X-ray images, that are preprocessed using an Auto-encoder before feeding them into the network. The proposed work includes a hybrid Ensemble approach for both datasets. The proposed approach uses three Well-known convolutional neural network (CNN) models for Ensemble. These models include MobilenetV2, ResNet152, and DenseNet169. These models were individually trained using transfer learning (as the models were Pre-trained on the ImageNet dataset) and fine-tuned. The results were compared with those of the proposed method. The Kermany pneumonia chest X-ray dataset results in the proposed Ensemble approach are as follows: We achieved 95.03% classification accuracy. The results of the proposed Ensemble approach on the Father Muller Hospital dataset are as follows: We achieved 82.60% classification accuracy. © 2024 Copyright held by the owner/author(s).Item Transfer learning based code-mixed part-of-speech tagging using character level representations for Indian languages(Springer Science and Business Media Deutschland GmbH, 2023) Anand Kumar, A.K.; Padannayil, S.K.Massive amounts of unstructured content have been generated day-by-day on social media platforms like Facebook, Twitter and blogs. Analyzing and extracting useful information from this vast amount of text content is a challenging process. Social media have currently provided extensive opportunities for researchers and practitioners to do adequate research on this area. Most of the text content in social media tend to be either in English or code-mixed regional languages. In a multilingual country like India, code-mixing is the usual fashion witnessed in social media discussions. Multilingual users frequently use Roman script, an convenient mode of expression, instead of the regional language script for posting messages on social media and often mix it with English into their native languages. Stylistic and grammatical irregularities are significant challenges in processing the code-mixed text using conventional methods. This paper explains the new word embedding via character level representation as features for POS tagging the code-mixed text in Indian languages using the ICON-2015, ICON-2016 NLP tools contest data set. The proposed word embedding features are context-appended, and the well-known Support Vector Machine (SVM) classifier has been used to train the system. We have combined the Facebook, Twitter, and WhatsApp code-mixed data of three Indian languages to train the Transfer learning based language-independent and source independent POS tagging. The experimental results demonstrated that the proposed transfer method achieved state-of-the-art accuracy in 12 systems out of 18 systems for the ICON data set. © 2021, The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature.Item Transfer Learning-Based Fault Diagnosis of Internal Combustion (IC) Engine Gearbox Using Radar Plots(John Wiley and Sons Ltd, 2024) S, S.; Srivatsan, B.; Sugumaran, V.; Ravikumar, K.N.; Kumar, H.; Mahamuni, V.S.Due to constant loads, gear wear, and harsh working conditions, gearboxes are subject to fault occurrences. Faults in the gearbox can cause damage to the engine components, create unnecessary noise, degrade efficiency, and impact power transfer. Hence, the detection of faults at an early stage is highly necessary. In this work, an effort was made to use transfer learning to identify gear failures under five gear conditions—healthy condition, 25% defect, 50% defect, 75% defect, and 100% defect—and three load conditions—no load, T1 = 9.6, and T2 = 13.3 Nm. Vibration signals were collected for various gear and load conditions using an accelerometer mounted on the casing of the gearbox. The load was applied using an eddy current dynamometer on the output shaft of the engine. The obtained vibration signals were processed and stored as vibration radar plots. Residual network (ResNet)-50, GoogLenet, Visual Geometry Group 16 (VGG-16), and AlexNet were the network models used for transfer learning in this study. Hyperparameters, including learning rate, optimizer, train-test split ratio, batch size, and epochs, were varied in order to achieve the highest classification accuracy for each pretrained network. From the results obtained, VGG-16 pretrained network outperformed all other networks with a classification accuracy of 100%. © © 2024 S. Naveen Venkatesh et al.Item Channel Pruning of Transfer Learning Models Using Novel Techniques(Institute of Electrical and Electronics Engineers Inc., 2024) Pragnesh, P.; Mohan, B.R.This research paper delves into the challenges associated with deep learning models, specifically focusing on transfer learning. Despite the effectiveness of widely used models such as VGGNet, ResNet, and GoogLeNet, their deployment on resource-constrained devices is impeded by high memory bandwidth and computational costs, and to overcome these limitations, the study proposes pruning as a viable solution. Numerous parameters, particularly in fully connected layers, contribute minimally to computational costs, so we focus on convolution layers' pruning. The research explores and evaluates three innovative pruning methods: the Max3 Saliency pruning method, the K-Means clustering algorithm, and the Singular Value Decomposition (SVD) approach. The Max3 Saliency pruning method introduces a slight variation by using the three maximum values of the kernel instead of all nine to compute the saliency score. This method is the most effective, substantially reducing parameter and Floating Point Operations (FLOPs) for both VGG16 and ResNet56 models. Notably, VGG16 demonstrates a remarkable 46.19% reduction in parameters and a 61.91% reduction in FLOPs. Using the Max3 Saliency pruning method, ResNet56 shows a 35.15% reduction in parameters and FLOPs. The K-Means pruning algorithm is also successful, resulting in a 40.00% reduction in parameters for VGG16 and a 49.20% reduction in FLOPs. In the case of ResNet56, the K-Means algorithm achieved a 31.01% reduction in both parameters and FLOPs. While the Singular Value Decomposition (SVD) approach provides a new set of values for condensed channels, its overall pruning ratio is smaller than the Max3 Saliency and K-Means methods. The SVD pruning method prunes 20.07% parameter reduction and a 24.64% reduction in FLOPs achieved for VGG16, along with a 16.94% reduction in both FLOPs and parameters for ResNet56. Compared with the state-of-the-art methods, the Max3 Saliency and K-Means pruning methods performed better in Flops reduction metrics. © 2024 The Authors.