Varun Gupta
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Varun Gupta has a total of 77 co-authors. Here are the top co-authors by number of shared publications:
- Jitender Kumar Chhabra
National Institute of Technology
7 shared publications Praveen Kumar
Department of Computer Science and Engineering, Chandigarh College of Engineering and Technology, Chandigarh, India
5 shared publicationsNitigya Sambyal
Department of Computer Science and Engineering, Punjab Engineering College (Deemed to be University), Chandigarh, India
4 shared publications- Prakash Chandra Chhipa
Luleå Tekniska Universitet
3 shared publications - Aaron Roth
University of Pennsylvania
2 shared publications - Deepika Vikas Agrawal
National Institute of Technical Teachers Training and Research
2 shared publications
Latest Publications
- Open Access
- 26 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 175676-175685
Autoregressive T-Cell-Receptors Algorithm-Based Interventions for Multivariate Cointegration Estimation and Forecast
Multivariate time series forecasting presents significant challenges due to non-stationarity, autocorrelation, and structural shifts, especially when latent long-term cointegration dynamics are present. This paper introduces the T-Cell Receptor (TCR) algorithm, a novel forecasting framework inspired by immune system selectivity and adaptability. Mimicking TCR antigen recognition functions, the model employs Orthogonal Least Squares (OLS) for dynamic regressor selection and Inter-Cell Cohesion Factors (ICCF) to quantifies each regressor’s contribution to error reduction. This enables adaptive identification of cointegrated lag structures, gaining an advantage over traditional models that rely on fixed heuristics, in turn, performing adaptive forecasting across volatile regimes. TCR was evaluated on separate samples of daily adjusted closing prices from the S&P 500 (SNP) and Infosys (INFY), representing datasets with differing volatility but evident cointegration. Benchmarking against ARIMA, SARIMA, Single Exponential Smoothing (SES), and the Dendritic Neuron Model (DNM), shows that TCR captures the fluctuations more effectively, particularly in high-volatility series, and adapts better than smoothing-based models. ACF analysis confirms stationarity, while error boxplots highlight predictive stability. Quantitatively, TCR achieved the lowest RMSE and MAE on volatile SNP data, with statistically significant improvements (p < 0.05) over SARIMA, SES, and DNM. The study validates TCR’s novelty as an interpretable, biologically inspired, and cointegration-aware forecasting model. Future extensions may address high-dimensional, realtime, and hybrid learning scenarios.
- 26 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 13, 175676-175685
Open AccessAutoregressive T-Cell-Receptors Algorithm-Based Interventions for Multivariate Cointegration Estimation and Forecast
Multivariate time series forecasting presents significant challenges due to non-stationarity, autocorrelation, and structural shifts, especially when latent long-term cointegration dynamics are present. This paper introduces the T-Cell Receptor (TCR) algorithm, a novel forecasting framework inspired by immune system selectivity and adaptability. Mimicking TCR antigen recognition functions, the model employs Orthogonal Least Squares (OLS) for dynamic regressor selection and Inter-Cell Cohesion Factors (ICCF) to quantifies each regressor’s contribution to error reduction. This enables adaptive identification of cointegrated lag structures, gaining an advantage over traditional models that rely on fixed heuristics, in turn, performing adaptive forecasting across volatile regimes. TCR was evaluated on separate samples of daily adjusted closing prices from the S&P 500 (SNP) and Infosys (INFY), representing datasets with differing volatility but evident cointegration. Benchmarking against ARIMA, SARIMA, Single Exponential Smoothing (SES), and the Dendritic Neuron Model (DNM), shows that TCR captures the fluctuations more effectively, particularly in high-volatility series, and adapts better than smoothing-based models. ACF analysis confirms stationarity, while error boxplots highlight predictive stability. Quantitatively, TCR achieved the lowest RMSE and MAE on volatile SNP data, with statistically significant improvements (p < 0.05) over SARIMA, SES, and DNM. The study validates TCR’s novelty as an interpretable, biologically inspired, and cointegration-aware forecasting model. Future extensions may address high-dimensional, realtime, and hybrid learning scenarios.
- 7 August 2025
- Published by Springer Nature
- p. 288-296
Contrastive Learning-Based Approach for Automated Road Damage Detection
- 7 August 2025
- Published by Springer Nature
- p. 288-296
Contrastive Learning-Based Approach for Automated Road Damage Detection
- Open Access
- 28 December 2024
- Published by Springer Nature in Health Information Science and Systems
- Vol. 13 (1) , 1-2
Correction: Margin-aware optimized contrastive learning for enhanced self-supervised histopathological image classification
[This corrects the article DOI: 10.1007/s13755-024-00316-4.].
- 28 December 2024
- Published by Springer Nature in Health Information Science and Systems
- Vol. 13 (1) , 1-2
Open AccessCorrection: Margin-aware optimized contrastive learning for enhanced self-supervised histopathological image classification
[This corrects the article DOI: 10.1007/s13755-024-00316-4.].
- 29 November 2024
- Published by Springer Nature in Health Information Science and Systems
- Vol. 13 (1) , 1-22
Margin-aware optimized contrastive learning for enhanced self-supervised histopathological image classification
Histopathological images, characterized by their high resolution and intricate cellular structures, present unique challenges for automated analysis. Traditional supervised learning-based methods often rely on extensive labeled datasets, which are labour-intensive and expensive. In learning representations, self-supervised learning techniques have shown promising outcomes directly from raw image data without manual annotations. In this paper, we propose a novel margin-aware optimized contrastive learning approach to enhance representation learning from histopathological images using a self-supervised approach. The proposed approach optimizes contrastive learning with a margin-based strategy to effectively learn discriminative representations while enforcing a semantic similarity threshold. In the proposed loss function, a margin is used to enforce a certain level of similarity between positive pairs in the embedding space, and a scaling factor is introduced to adjust the sensitivity of the loss, thereby enhancing the discriminative capacity of the learned representations. Our approach demonstrates robust generalization in in- and out-domain settings through comprehensive experimental evaluations conducted on five distinct benchmark histopathological datasets belonging to three cancer types. The results obtained on different experimental settings show that the proposed approach outmatched the state-of-the-art approaches in cross-domain and cross-disease settings.
- 29 November 2024
- Published by Springer Nature in Health Information Science and Systems
- Vol. 13 (1) , 1-22
Margin-aware optimized contrastive learning for enhanced self-supervised histopathological image classification
Histopathological images, characterized by their high resolution and intricate cellular structures, present unique challenges for automated analysis. Traditional supervised learning-based methods often rely on extensive labeled datasets, which are labour-intensive and expensive. In learning representations, self-supervised learning techniques have shown promising outcomes directly from raw image data without manual annotations. In this paper, we propose a novel margin-aware optimized contrastive learning approach to enhance representation learning from histopathological images using a self-supervised approach. The proposed approach optimizes contrastive learning with a margin-based strategy to effectively learn discriminative representations while enforcing a semantic similarity threshold. In the proposed loss function, a margin is used to enforce a certain level of similarity between positive pairs in the embedding space, and a scaling factor is introduced to adjust the sensitivity of the loss, thereby enhancing the discriminative capacity of the learned representations. Our approach demonstrates robust generalization in in- and out-domain settings through comprehensive experimental evaluations conducted on five distinct benchmark histopathological datasets belonging to three cancer types. The results obtained on different experimental settings show that the proposed approach outmatched the state-of-the-art approaches in cross-domain and cross-disease settings.
- 5 September 2024
- Published by Springer Nature in Archives of Computational Methods in Engineering
- Vol. 32 (2) , 1199-1227
Preserving Artistic Heritage: A Comprehensive Review of Virtual Restoration Methods for Damaged Artworks
- 5 September 2024
- Published by Springer Nature in Archives of Computational Methods in Engineering
- Vol. 32 (2) , 1199-1227
Preserving Artistic Heritage: A Comprehensive Review of Virtual Restoration Methods for Damaged Artworks
- Open Access
Functional Knowledge Transfer with Self-supervised Representation Learning
This work investigates the unexplored usability of self-supervised representation learning in the direction of functional knowledge transfer. In this work, functional knowledge transfer is achieved by joint optimization of self-supervised learning pseudo task and supervised learning task, improving supervised learning task performance. Recent progress in self-supervised learning uses a large volume of data, which becomes a constraint for its applications on small-scale datasets. This work shares a simple yet effective joint training framework that reinforces human-supervised task learning by learning self-supervised representations just-in-time and vice versa. Experiments on three public datasets from different visual domains, Intel Image, CIFAR, and APTOS, reveal a consistent track of performance improvements on classification tasks during joint optimization. Qualitative analysis also supports the robustness of learnt representations. Source code and trained models are available on GitHub.
Open AccessFunctional Knowledge Transfer with Self-supervised Representation Learning
This work investigates the unexplored usability of self-supervised representation learning in the direction of functional knowledge transfer. In this work, functional knowledge transfer is achieved by joint optimization of self-supervised learning pseudo task and supervised learning task, improving supervised learning task performance. Recent progress in self-supervised learning uses a large volume of data, which becomes a constraint for its applications on small-scale datasets. This work shares a simple yet effective joint training framework that reinforces human-supervised task learning by learning self-supervised representations just-in-time and vice versa. Experiments on three public datasets from different visual domains, Intel Image, CIFAR, and APTOS, reveal a consistent track of performance improvements on classification tasks during joint optimization. Qualitative analysis also supports the robustness of learnt representations. Source code and trained models are available on GitHub.
- Open Access
- 4 July 2023
- Published by EDP Sciences in ITM Web of Conferences
- Vol. 54, 01013
Artwork restoration using paired image translation-based generative adversarial networks
Preservation of the artworks has historical and cultural importance. However, with time, environmental factors severely affect artworks, and these damages are often complicated to repair manually and through traditional methods. We propose a method to restore artwork that has been damaged over time. This work proposes a systematic approach using paired image-to-image translation based on a generative adversarial network. The experimental results have been quantitatively evaluated. The experimental results obtained from the presented work visually prove that the presented approach of artwork restoration completely restores the damaged artwork.
- 4 July 2023
- Published by EDP Sciences in ITM Web of Conferences
- Vol. 54, 01013
Open AccessArtwork restoration using paired image translation-based generative adversarial networks
Preservation of the artworks has historical and cultural importance. However, with time, environmental factors severely affect artworks, and these damages are often complicated to repair manually and through traditional methods. We propose a method to restore artwork that has been damaged over time. This work proposes a systematic approach using paired image-to-image translation based on a generative adversarial network. The experimental results have been quantitatively evaluated. The experimental results obtained from the presented work visually prove that the presented approach of artwork restoration completely restores the damaged artwork.
- Open Access
Learning Self-Supervised Representations for Label Efficient Cross-Domain Knowledge Transfer on Diabetic Retinopathy Fundus Images
This work presents a novel label-efficient selfsupervised representation learning-based approach for classifying diabetic retinopathy (DR) images in cross-domain settings. Most of the existing DR image classification methods are based on supervised learning which requires a lot of time-consuming and expensive medical domain experts-annotated data for training. The proposed approach uses the prior learning from the source DR image dataset to classify images drawn from the target datasets. The image representations learned from the unlabeled source domain dataset through contrastive learning are used to classify DR images from the target domain dataset. Moreover, the proposed approach requires a few labeled images to perform successfully on DR image classification tasks in cross-domain settings. The proposed work experiments with four publicly available datasets: EyePACS, APTOS 2019, MESSIDOR-I, and Fundus Images for self-supervised representation learning-based DR image classification in cross-domain settings. The proposed method achieves state-of-the-art results on binary and multiclassification of DR images, even in cross-domain settings. The proposed method outperforms the existing DR image binary and multi-class classification methods proposed in the literature. The proposed method is also validated qualitatively using class activation maps, revealing that the method can learn explainable image representations. The source code and trained models are published on GitHub.
Open AccessLearning Self-Supervised Representations for Label Efficient Cross-Domain Knowledge Transfer on Diabetic Retinopathy Fundus Images
This work presents a novel label-efficient selfsupervised representation learning-based approach for classifying diabetic retinopathy (DR) images in cross-domain settings. Most of the existing DR image classification methods are based on supervised learning which requires a lot of time-consuming and expensive medical domain experts-annotated data for training. The proposed approach uses the prior learning from the source DR image dataset to classify images drawn from the target datasets. The image representations learned from the unlabeled source domain dataset through contrastive learning are used to classify DR images from the target domain dataset. Moreover, the proposed approach requires a few labeled images to perform successfully on DR image classification tasks in cross-domain settings. The proposed work experiments with four publicly available datasets: EyePACS, APTOS 2019, MESSIDOR-I, and Fundus Images for self-supervised representation learning-based DR image classification in cross-domain settings. The proposed method achieves state-of-the-art results on binary and multiclassification of DR images, even in cross-domain settings. The proposed method outperforms the existing DR image binary and multi-class classification methods proposed in the literature. The proposed method is also validated qualitatively using class activation maps, revealing that the method can learn explainable image representations. The source code and trained models are published on GitHub.