TEJASWINI KAR

Dr. TEJASWINI KAR

KIIT University, Bhubaneswar, India
orcid

Research Focus & Metrics

Academic Career Length

12Years (2015-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
DSDebaditya Shome
Debaditya Shome
TKTEJASWINI KAR
TEJASWINI KAR
Sachi Nandan MohantySachi Nandan Mohanty
Sachi Nandan Mohanty
Prayag TiwariPrayag Tiwari
Prayag Tiwari
Khan MuhammadKhan Muhammad
Khan Muhammad
Abdul Khader Jilani SaudagarAbdul Khader Jilani Saudagar
Abdul Khader Jilani Saudagar
In the recent pandemic, accurate and rapid testing of patients remained a critical task in the diagnosis and control of COVID-19 disease spread in the healthcare industry. Because of the sudden increase in cases, most countries have faced scarcity and a low rate of testing. Chest X-rays have been shown in the literature to be a potential source of testing for COVID-19 patients, but manually checking X-ray reports is time-consuming and error-prone. Considering these limitations and the advancements in data science, we proposed a Vision Transformer-based deep learning pipeline for COVID-19 detection from chest X-ray-based imaging. Due to the lack of large data sets, we collected data from three open-source data sets of chest X-ray images and aggregated them to form a 30 K image data set, which is the largest publicly available collection of chest X-ray images in this domain to our knowledge. Our proposed transformer model effectively differentiates COVID-19 from normal chest X-rays with an accuracy of 98% along with an AUC score of 99% in the binary classification task. It distinguishes COVID-19, normal, and pneumonia patient’s X-rays with an accuracy of 92% and AUC score of 98% in the Multi-class classification task. For evaluation on our data set, we fine-tuned some of the widely used models in literature, namely, EfficientNetB0, InceptionV3, Resnet50, MobileNetV3, Xception, and DenseNet-121, as baselines. Our proposed transformer model outperformed them in terms of all metrics. In addition, a Grad-CAM based visualization is created which makes our approach interpretable by radiologists and can be used to monitor the progression of the disease in the affected lungs, assisting healthcare.
1

30 March 2024 in Artificial Intelligence Review
Verified
Journal Article
Open access
PDF file
TKTEJASWINI KAR
TEJASWINI KAR
PKP. Kanungo
P. Kanungo
Sachi Nandan MohantySachi Nandan Mohanty
Sachi Nandan Mohanty
Sven GroppeSven Groppe
Sven Groppe
JGJinghua Groppe
Jinghua Groppe
The integration of high data transmission rates and the recent digital multimedia technology, paves the way to access a huge amount of video over the internet, in seconds. Additionally, uploading videos to different websites is no more confined to expert software professionals resulting in duplication of video data which led to exorbitant growth of multimedia information in cyberspace in a short span of time. This necessitates the development of efficient data management techniques including storage, searching and annotation mechanism. Automatic shot boundary detection is considered to be the first and foremost step towards such management. It is a booming area of research gaining attention in the domain of image processing, computer vision and pattern recognition. In this review paper, we present a detailed description of the methods and algorithms of shot boundary detection, reported in the last two decades. This review shows that using multiple features performs well in comparison to using only a single feature in the shot boundary detection problem although it leads to higher complexity. The major sources of disturbance in the boundary detection are the sudden illumination variation and presence of high motion in the video. An adaptive threshold outperforms a single global threshold in the boundary detection problem and the threshold requirement can be avoided through learning based strategies at the cost of larger training data and higher computation time. Moreover the present review includes a critical analysis of relative merits and demerits of existing algorithms and finally opens promising research directions in the area.

22 August 2024 in IEEE Access
Verified
Journal Article
Open access
TPTapaswini Pattnaik
Tapaswini Pattnaik
PKPriyadarshi Kanungo
Priyadarshi Kanungo
UPUpasana Kumar Panigrahi
Upasana Kumar Panigrahi
TKTEJASWINI KAR
TEJASWINI KAR
Prince JainPrince Jain
Prince Jain
MIMohammad Tariqul Islam
Mohammad Tariqul Islam
The most popular technique for converting two-class images into binary images is thresholding. However, thresholding methods tend to perform poorly when dealing with images affected by uneven lighting. To address this issue, local thresholding techniques are commonly used. While pixel-based local thresholding methods can achieve high accuracy, they are computationally complex. Window-based local thresholding presents challenges in selecting the initial window and determining the criterion function for dividing the image into smaller versions. In this study, a novel technique is proposed to improve the effectiveness of binarizing images with uneven lighting. The proposed method is based on a low-complexity functional neural network model (LC-FLANN) to estimate an image’s illumination surface. The effectiveness of the proposed technique has been evaluated using five widely used uneven lighting image binarization techniques and various uneven light image variations. The results show that the proposed approach outperforms other alternatives in both qualitative and quantitative metrics. It achieved an average F-Measure score of 0.97, a Jaccard Index (JI) score of 0.95, and a Percentage of Misclassification Error (PME) 1.42%, demonstrating superior overall performance.

18 March 2017 in Signal, Image and Video Processing
Verified
Journal Article
Closed access
PDF file
TKTEJASWINI KAR
TEJASWINI KAR
PKP. Kanungo
P. Kanungo

21 September 2023 in ICST Transactions on Scalable Information Systems
Verified
Journal Article
Open access
PDF file
SASnehaprava Acharya
Snehaprava Acharya
TKTEJASWINI KAR
TEJASWINI KAR
USUmesh Chandra Samal
Umesh Chandra Samal
PPPrasant Kumar Patra
Prasant Kumar Patra
INTRODUCTION: Automatic detection of rice plant diseases at early stage from its images is quite beneficial over traditional verification methods. OBJECTIVES: Recent years machine learning (ML) approaches are more efficient in disease classification task. In current generation the statistical machine learning algorithm which shows state-of-arts performance is Support Vector Machine (SVM) and variants of SVM. METHODS: SVM has an excellent learning performance for linear and non-linear data samples. It works for Quadratic Programming Problems (QPP) due to which it has the drawback of computational complexity. However QPP can be solved linearly with the help of Least Square SVM(LS-SVM) approach. In LS-SVM the epsilon tube and slack variables of SVM are replaced with error variables. The distance is calculated by error square value. RESULTS: In this research performance comparison is made between SVM and LS-SVM for rice leaf diseases such as Bacterial Leaf Blight (BLB), Brown spot(BS), Leaf smut(LS) and Leaf Blast using two datasets (DS1 and DS2).Accuracy of LS-SVM is found to be 91.3% and 98.87% for DS1 and DS2 respectively whereas accuracy of SVM is 83.3% and 98.75% for DS1 and DS2 respectively. CONCLUSION: Performance of LS-SVM outperformed than SVM in terms of accuracy.

31 December 2023 in e-Prime - Advances in Electrical Engineering, Electronics and Energy
Verified
Journal Article
Open access
TPTapaswini Pattnaik
Tapaswini Pattnaik
PKPriyadarshi Kanungo
Priyadarshi Kanungo
TKTEJASWINI KAR
TEJASWINI KAR
UPUpasana Kumar Panigrahi
Upasana Kumar Panigrahi