Shrutika S. Sawant
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Shrutika S. Sawant has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Manoharan Prabukumar
School of Computer Science Engineering and Information Systems, Vellore Institute of Technology, Vellore, India
5 shared publications - Sathishkumar Samiappan
Department of Biosystems Engineering & Soil Science, University of Tennessee at Knoxville, Knoxville, TN 37996, USA
2 shared publications Agilandeeswari Loganathan
School of Information Technology & Engineering, Vellore Institute of Technology, Vellore, India
1 shared publication- F. X. Erick
Fraunhofer Institute for Integrated Circuits IIS, Germany
1 shared publication - Farhan A. Alenizi
Department of Electrical Engineering, College of Engineering, Prince Sattam bin Abdulaziz University, Al- Kharj, Saudi Arabia
1 shared publication - Nina Holzer
Fraunhofer Institute for Integrated Circuits IIS, Germany
1 shared publication
Latest Publications
- Open Access
- 21 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Self-supervised representation learning using multimodal Transformer for emotion recognition
In this paper, we present a Modality-Agnostic Transformer based Self-Supervised Learning (MATS2L) for emotion recognition using physiological signals. The proposed approach consists of two stages: a) Pretext stage, where the transformer model is pre-trained with unlabeled physiological signal data using masked signal prediction as pre-training task and form contextualized signal representations. b) Downstream stage, where self-supervised learning (SSL) representations extracted from a pre-trained model are utilized for emotion recognition tasks. Modality-agnostic approach allows the transformer model to focus on exploring mutual features among different physiological signals and learning more meaningful embeddings to estimate emotions effectively. We conduct several experiments on a public dataset WESAD and perform comparisons with fully supervised and other competitive SSL approaches. Experimental results showed that the proposed approach is capable of learning meaningful features and superior to other competitive SSL approaches. Moreover, a transformer model trained on SSL features outperforms fully supervised transformer model. We also present detailed ablation studies to prove the robustness of our approach.
- 21 September 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessSelf-supervised representation learning using multimodal Transformer for emotion recognition
In this paper, we present a Modality-Agnostic Transformer based Self-Supervised Learning (MATS2L) for emotion recognition using physiological signals. The proposed approach consists of two stages: a) Pretext stage, where the transformer model is pre-trained with unlabeled physiological signal data using masked signal prediction as pre-training task and form contextualized signal representations. b) Downstream stage, where self-supervised learning (SSL) representations extracted from a pre-trained model are utilized for emotion recognition tasks. Modality-agnostic approach allows the transformer model to focus on exploring mutual features among different physiological signals and learning more meaningful embeddings to estimate emotions effectively. We conduct several experiments on a public dataset WESAD and perform comparisons with fully supervised and other competitive SSL approaches. Experimental results showed that the proposed approach is capable of learning meaningful features and superior to other competitive SSL approaches. Moreover, a transformer model trained on SSL features outperforms fully supervised transformer model. We also present detailed ablation studies to prove the robustness of our approach.
- 3 June 2022
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 43 (11) , 3990-4024
Multi-objective multi-verse optimizer based unsupervised band selection for hyperspectral image classification
Hyperspectral band selection is one of the efficacious ways to diminish the size of hyperspectral images. The process of selecting a few useful bands will be successful when two fundamental aspects are considered: information abundance and redundancy among the chosen bands. However, selecting the suitable number of bands in an ill-posed classification problem remains challenging. Overcoming this issue, a novel unsupervised multi-objective multi-verse optimizer-based band selection (MOMVOBS) approach is proposed. It explores optimal trade-offs among the different traits of the objective functions namely information richness, less redundancy and the number of bands to be selected. These three objective functions are optimized simultaneously using a multiverse optimizer (MVO) to obtain the best solutions. To evaluate the quality of selected bands, two widely used supervised classifiers are used, such as support vector machine (SVM) and K-nearest neighbour (KNN). Experimental results evidence for the superiority of the proposed approach over the recent multi-objective optimization-based band selection approaches by selecting the highly informative distinct bands that have better classification performance on four benchmark hyperspectral data sets. The proposed MOMVOBS have obtained 79.50% and 71.35% of overall accuracy for SVM and KNN classifier, respectively, in Indian Pines dataset with 10% of band retention, 93.06% and 88.88% of overall accuracy for SVM and KNN classifier, respectively, in Salinas dataset with 10% of band retention, 92.86% and 85.35% of overall accuracy for SVM and KNN classifier, respectively, in Pavia University dataset with 15% band retention, and 92.42% and 85.33% of overall accuracy for SVM and KNN classifier, respectively, in Botswana dataset with 11% band retention. The achievement of higher accuracy at less than 15% bands is significant.
- 3 June 2022
- Published by Taylor & Francis in International Journal of Remote Sensing
- Vol. 43 (11) , 3990-4024
Multi-objective multi-verse optimizer based unsupervised band selection for hyperspectral image classification
Hyperspectral band selection is one of the efficacious ways to diminish the size of hyperspectral images. The process of selecting a few useful bands will be successful when two fundamental aspects are considered: information abundance and redundancy among the chosen bands. However, selecting the suitable number of bands in an ill-posed classification problem remains challenging. Overcoming this issue, a novel unsupervised multi-objective multi-verse optimizer-based band selection (MOMVOBS) approach is proposed. It explores optimal trade-offs among the different traits of the objective functions namely information richness, less redundancy and the number of bands to be selected. These three objective functions are optimized simultaneously using a multiverse optimizer (MVO) to obtain the best solutions. To evaluate the quality of selected bands, two widely used supervised classifiers are used, such as support vector machine (SVM) and K-nearest neighbour (KNN). Experimental results evidence for the superiority of the proposed approach over the recent multi-objective optimization-based band selection approaches by selecting the highly informative distinct bands that have better classification performance on four benchmark hyperspectral data sets. The proposed MOMVOBS have obtained 79.50% and 71.35% of overall accuracy for SVM and KNN classifier, respectively, in Indian Pines dataset with 10% of band retention, 93.06% and 88.88% of overall accuracy for SVM and KNN classifier, respectively, in Salinas dataset with 10% of band retention, 92.86% and 85.35% of overall accuracy for SVM and KNN classifier, respectively, in Pavia University dataset with 15% band retention, and 92.42% and 85.33% of overall accuracy for SVM and KNN classifier, respectively, in Botswana dataset with 11% band retention. The achievement of higher accuracy at less than 15% bands is significant.
- 9 September 2020
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 80 (2) , 1725-1748
A hybrid optimization approach for hyperspectral band selection based on wind driven optimization and modified cuckoo search optimization
- 9 September 2020
- Published by Springer Nature in Multimedia Tools and Applications
- Vol. 80 (2) , 1725-1748
A hybrid optimization approach for hyperspectral band selection based on wind driven optimization and modified cuckoo search optimization
- Open Access
- 22 June 2020
- Published by IM Publications Open LLP in Journal of Spectral Imaging
- Vol. 9
A modified Cuckoo Search algorithm based optimal band subset selection approach for hyperspectral image classification
Band selection is an effective way to reduce the size of hyperspectral data and to overcome the “curse of dimensionality” in ground object classification. This paper presents a band selection approach based on modified Cuckoo Search (CS) optimisation with correlation-based initialisation. CS is a popular metaheuristic algorithm with efficient optimisation capabilities for band selection. However, it can easily fall into local optimum solutions. To avoid falling into a local optimum, an initialisation strategy based on correlation is adopted instead of random initialisation to initiate the location of nests. Experimental results with Indian Pines, Salinas and Pavia University datasets show that the proposed approach obtains overall accuracy of 82.83 %, 94.83 % and 91.79 %, respectively, which is higher than the original CS algorithm, Genetic Algorithm (GA), Particle Swarm Optimisation (PSO) and Gray Wolf Optimisation (GWO).
- 22 June 2020
- Published by IM Publications Open LLP in Journal of Spectral Imaging
- Vol. 9
Open AccessA modified Cuckoo Search algorithm based optimal band subset selection approach for hyperspectral image classification
Band selection is an effective way to reduce the size of hyperspectral data and to overcome the “curse of dimensionality” in ground object classification. This paper presents a band selection approach based on modified Cuckoo Search (CS) optimisation with correlation-based initialisation. CS is a popular metaheuristic algorithm with efficient optimisation capabilities for band selection. However, it can easily fall into local optimum solutions. To avoid falling into a local optimum, an initialisation strategy based on correlation is adopted instead of random initialisation to initiate the location of nests. Experimental results with Indian Pines, Salinas and Pavia University datasets show that the proposed approach obtains overall accuracy of 82.83 %, 94.83 % and 91.79 %, respectively, which is higher than the original CS algorithm, Genetic Algorithm (GA), Particle Swarm Optimisation (PSO) and Gray Wolf Optimisation (GWO).
- Open Access
- 16 June 2020
- Published by IM Publications Open LLP in Journal of Spectral Imaging
- Vol. 9
A survey of band selection techniques for hyperspectral image classification
Hyperspectral images usually contain hundreds of contiguous spectral bands, which can precisely discriminate the various spectrally similar classes. However, such high-dimensional data also contain highly correlated and irrelevant information, leading to the curse of dimensionality (also called the Hughes phenomenon). It is necessary to reduce these bands before further analysis, such as land cover classification and target detection. Band selection is an effective way to reduce the size of hyperspectral data and to overcome the curse of the dimensionality problem in ground object classification. Focusing on the classification task, this article provides an extensive and comprehensive survey on band selection techniques describing the categorisation of methods, methodology used, different searching approaches and various technical difficulties, as well as their performances. Our purpose is to highlight the progress attained in band selection techniques for hyperspectral image classification and to identify possible avenues for future work, in order to achieve better performance in real-time operation.
- 16 June 2020
- Published by IM Publications Open LLP in Journal of Spectral Imaging
- Vol. 9
Open AccessA survey of band selection techniques for hyperspectral image classification
Hyperspectral images usually contain hundreds of contiguous spectral bands, which can precisely discriminate the various spectrally similar classes. However, such high-dimensional data also contain highly correlated and irrelevant information, leading to the curse of dimensionality (also called the Hughes phenomenon). It is necessary to reduce these bands before further analysis, such as land cover classification and target detection. Band selection is an effective way to reduce the size of hyperspectral data and to overcome the curse of the dimensionality problem in ground object classification. Focusing on the classification task, this article provides an extensive and comprehensive survey on band selection techniques describing the categorisation of methods, methodology used, different searching approaches and various technical difficulties, as well as their performances. Our purpose is to highlight the progress attained in band selection techniques for hyperspectral image classification and to identify possible avenues for future work, in order to achieve better performance in real-time operation.