S. Kalaivani
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
S. Kalaivani has a total of 30 co-authors. Here are the top co-authors by number of shared publications:
V. Sangeetha
School of Computer Science Engineering and Information Systems, Vellore Institute of Technology
3 shared publications- Abdulla Al Kafy
Department of Geography & the Environment, The University of Texas at Austin, Austin, USA
2 shared publications M. Suriya Begum
School of Computer Science Engineering and Information Systems, VIT
2 shared publications- Zullyadini A. Rahaman
Department of Geography & Environment, Faculty of Human Sciences, Sultan Idris Education University, Tanjung Malim, Malaysia
2 shared publications Abdul Fattah
Department of Geography, Florida State University, Tallahassee, FL, 32306, USA; Department of Urban and Regional Planning, Khulna University of Engineering and Technology, Khulna, Bangladesh
1 shared publication- Abdul Fattah
Department of Geography, Florida State University, Tallahassee, USA
1 shared publication
Latest Publications
- Open Access
- 3 November 2026
- Published by Frontiers Media SA in Frontiers in Bioinformatics
- Vol. 6, 1901330
Multi-color space image enhancement with weighted fusion and lightweight deep learning framework for AI-based breast cancer histopathology image classification
Introduction: Variations in Hematoxylin and Eosin (H&E) staining, illumination, tissue preparation, and image acquisition affect breast histopathology image analysis, altering nuclear contrast, color consistency, and downstream feature extraction. Separately, prior BreakHis studies have often used image-wise splitting, which can place a patient’s images in both training and test sets and inflate performance estimates. Methods: To address these two issues, this study introduced a Multi-color Space Image Enhancement (MSIE) approach and evaluates it under strict, leakage-controlled, patient-wise five-fold cross-validation on the BreakHis dataset. MSIE combines H&E color deconvolution, stain normalization, hematoxylin contrast processing with Contrast Limited Adaptive Histogram Equalization (CLAHE), Hue-Saturation-Value (HSV) space processing, eosin reconstruction, and H&E-weighted fusion. Before model training, the dataset underwent a decoded-pixel Secure Hash Algorithm 256-bit (SHA-256) integrity check, and 125 duplicate images were removed, reducing the dataset from 7,909 to 7,784. The Lightweight Convolutional Neural Network (LCNN) model is trained and tested on MSIE and original RGB images using an identical architecture and patient partitions, and MobileNetV2, ResNet50, DenseNet169, and EfficientNetB0 are evaluated under the same protocol to assess efficiency-accuracy trade-offs across model capacities. Results: Performance is assessed at the image and patient levels using accuracy, precision, recall, F1 Score, Receiver Operating Characteristic Area Under the Curve (ROC AUC), and Precision-Recall Area Under the Curve (PR AUC), alongside Fusion-coefficient sensitivity, magnification-specific performance, computational efficiency, Grad-CAM visualization, and independent BACH testing. At the patient level, original RGB achieved an accuracy of 0.8397 ± 0.0325 and F1-score of 0.8789 ± 0.0326, versus a mean accuracy of 0.7522 ± 0.2218 and F1-score of 0.7632 ± 0.2734 for MSIE. MSIE, however, produced a slightly higher mean ROC-AUC (0.9024 ± 0.1274 versus 0.8952 ± 0.1023). Higher-capacity ImageNet-pretrained architectures performed better on MSIE inputs, with EfficientNetB0 achieving the highest mean patient-level accuracy (0.9007 ± 0.0718) and ROC-AUC (0.9707 ± 0.0354). Independent evaluation on the BACH dataset using frozen BreaKHis-trained MSIE-LCNN models yielded an accuracy of 0.7056 and an ROC-AUC of 0.7012, indicating substantial cross-dataset domain shift. Discussion: Overall, MSIE provides a stain-aware representation whose downstream benefits are not universal and depend on the classifier architecture, fusion balance, operating point, and dataset domain, thereby supporting leakage-controlled and externally validated assessment of histopathology processing methods rather than claims of immediate clinical translation.
- 3 November 2026
- Published by Frontiers Media SA in Frontiers in Bioinformatics
- Vol. 6, 1901330
Open AccessMulti-color space image enhancement with weighted fusion and lightweight deep learning framework for AI-based breast cancer histopathology image classification
Introduction: Variations in Hematoxylin and Eosin (H&E) staining, illumination, tissue preparation, and image acquisition affect breast histopathology image analysis, altering nuclear contrast, color consistency, and downstream feature extraction. Separately, prior BreakHis studies have often used image-wise splitting, which can place a patient’s images in both training and test sets and inflate performance estimates. Methods: To address these two issues, this study introduced a Multi-color Space Image Enhancement (MSIE) approach and evaluates it under strict, leakage-controlled, patient-wise five-fold cross-validation on the BreakHis dataset. MSIE combines H&E color deconvolution, stain normalization, hematoxylin contrast processing with Contrast Limited Adaptive Histogram Equalization (CLAHE), Hue-Saturation-Value (HSV) space processing, eosin reconstruction, and H&E-weighted fusion. Before model training, the dataset underwent a decoded-pixel Secure Hash Algorithm 256-bit (SHA-256) integrity check, and 125 duplicate images were removed, reducing the dataset from 7,909 to 7,784. The Lightweight Convolutional Neural Network (LCNN) model is trained and tested on MSIE and original RGB images using an identical architecture and patient partitions, and MobileNetV2, ResNet50, DenseNet169, and EfficientNetB0 are evaluated under the same protocol to assess efficiency-accuracy trade-offs across model capacities. Results: Performance is assessed at the image and patient levels using accuracy, precision, recall, F1 Score, Receiver Operating Characteristic Area Under the Curve (ROC AUC), and Precision-Recall Area Under the Curve (PR AUC), alongside Fusion-coefficient sensitivity, magnification-specific performance, computational efficiency, Grad-CAM visualization, and independent BACH testing. At the patient level, original RGB achieved an accuracy of 0.8397 ± 0.0325 and F1-score of 0.8789 ± 0.0326, versus a mean accuracy of 0.7522 ± 0.2218 and F1-score of 0.7632 ± 0.2734 for MSIE. MSIE, however, produced a slightly higher mean ROC-AUC (0.9024 ± 0.1274 versus 0.8952 ± 0.1023). Higher-capacity ImageNet-pretrained architectures performed better on MSIE inputs, with EfficientNetB0 achieving the highest mean patient-level accuracy (0.9007 ± 0.0718) and ROC-AUC (0.9707 ± 0.0354). Independent evaluation on the BACH dataset using frozen BreaKHis-trained MSIE-LCNN models yielded an accuracy of 0.7056 and an ROC-AUC of 0.7012, indicating substantial cross-dataset domain shift. Discussion: Overall, MSIE provides a stain-aware representation whose downstream benefits are not universal and depend on the classifier architecture, fusion balance, operating point, and dataset domain, thereby supporting leakage-controlled and externally validated assessment of histopathology processing methods rather than claims of immediate clinical translation.
- Open Access
- 30 October 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1964967
MSERNet: a multi-scale edge-guided network for JPEG compression artifact restoration in retrospective ovarian ultrasound images
Introduction: JPEG compression is widely used to store and transmit medical ultrasound images but can introduce artifacts that affect image quality and subsequent image analysis. This study proposes a Multi-Scale Edge Guided Restoration Network (MSERNet) for reducing JPEG compression artifacts while preserving tissue textures and anatomical boundaries in ultrasound images. Methods: An ethically approved, anonymized private ovarian ultrasound dataset was curated and used to generate JPEG-compressed and recompressed images at multiple quality levels. MSERNet integrates multi-scale feature extraction, adaptive feature fusion, and edge refinement within a residual learning framework. The model was evaluated on the private dataset and an independent public ultrasound dataset using the Structural Similarity Index (SSIM), Peak Signal-to-Noise Ratio (PSNR), and Visual Information Fidelity (VIF). Comparative experiments with existing restoration methods, ablation analysis, repeated-run evaluation, statistical analysis, and computational analysis were performed. Cross-dataset evaluation assessed generalization without additional fine-tuning. A downstream binary classification experiment using a transfer learning-based ResNet50 classifier was conducted on restored images, with performance assessed using accuracy, precision, sensitivity, specificity, F1-score, and area under the curve (AUC). Results: MSERNet reduced JPEG compression artifacts and improved image quality across the evaluated compression levels and datasets, with quantitative metrics supported by visual assessment. Comparative and ablation experiments assessed the method’s performance and component contributions, while repeated-run evaluation examined consistency. Cross-dataset evaluation demonstrated stable restoration performance without additional fine-tuning. The downstream classification experiment indicated that the restored images retained information relevant to cystic-versus-normal classification. Discussion: These findings highlight the potential of MSERNet to improve the quality of JPEG-compressed ultrasound images for subsequent analysis. Further validation on larger and more diverse datasets is needed to assess its generalizability and applicability to other medical image-analysis tasks.
- 30 October 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1964967
Open AccessMSERNet: a multi-scale edge-guided network for JPEG compression artifact restoration in retrospective ovarian ultrasound images
Introduction: JPEG compression is widely used to store and transmit medical ultrasound images but can introduce artifacts that affect image quality and subsequent image analysis. This study proposes a Multi-Scale Edge Guided Restoration Network (MSERNet) for reducing JPEG compression artifacts while preserving tissue textures and anatomical boundaries in ultrasound images. Methods: An ethically approved, anonymized private ovarian ultrasound dataset was curated and used to generate JPEG-compressed and recompressed images at multiple quality levels. MSERNet integrates multi-scale feature extraction, adaptive feature fusion, and edge refinement within a residual learning framework. The model was evaluated on the private dataset and an independent public ultrasound dataset using the Structural Similarity Index (SSIM), Peak Signal-to-Noise Ratio (PSNR), and Visual Information Fidelity (VIF). Comparative experiments with existing restoration methods, ablation analysis, repeated-run evaluation, statistical analysis, and computational analysis were performed. Cross-dataset evaluation assessed generalization without additional fine-tuning. A downstream binary classification experiment using a transfer learning-based ResNet50 classifier was conducted on restored images, with performance assessed using accuracy, precision, sensitivity, specificity, F1-score, and area under the curve (AUC). Results: MSERNet reduced JPEG compression artifacts and improved image quality across the evaluated compression levels and datasets, with quantitative metrics supported by visual assessment. Comparative and ablation experiments assessed the method’s performance and component contributions, while repeated-run evaluation examined consistency. Cross-dataset evaluation demonstrated stable restoration performance without additional fine-tuning. The downstream classification experiment indicated that the restored images retained information relevant to cystic-versus-normal classification. Discussion: These findings highlight the potential of MSERNet to improve the quality of JPEG-compressed ultrasound images for subsequent analysis. Further validation on larger and more diverse datasets is needed to assess its generalizability and applicability to other medical image-analysis tasks.
- Open Access
- 22 September 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1934912
Frequency response compensation to optimize pretrained deep learning models in lung nodule malignancy classification
Purpose: Accurate differentiation of benign and malignant pulmonary nodules on computed tomography (CT) is important for early lung-cancer assessment. However, classification remains challenging because benign and malignant nodules may exhibit subtle differences in texture, margin irregularity, boundary sharpness, and internal morphology that may not be fully represented by conventional convolutional feature extraction. Methods: A frequency response Bilinear Lag-Lead (BT-LL) EfficientNetB3 framework was developed for benign-malignant pulmonary nodule classification. The proposed approach integrates Lag-Lead frequency compensation into the pretrained feature-extraction process to refine intermediate representations. The Lag component performs structural smoothing to preserve low-frequency information, whereas the Lead component enhances high-frequency boundary and texture characteristics. These corresponding representations are integrated with the original feature maps through a dedicated fusion operation to obtain frequency-refined features for classification. The compensation filters are fixed, thereby avoiding additional parameter learning within the filtering operation. Patient-level data partitioning was used to ensure that nodules from the same patient were not distributed across different experimental subsets, reducing the risk of information leakage. Results: The proposed model achieved an accuracy, sensitivity, and specificity of 0.90, an F1-score of 0.81, and an AUROC of 0.95 on the held-out patient-level test set. Performance was additionally assessed using patient-level cross-validation, cross imbalance analysis, an IQ-OTH/NCCD image-level evaluation and an independent BUS-BRA patient-level evaluation. The experimental analysis further included statistical significance testing using McNemar’s test for pairwise model comparisons, followed by the Holm correction to control for multiple comparisons, while Cochran’s Q test was employed to assess performance differences across multiple model configurations. The combined configuration achieved the best sensitivity and balanced accuracy relative to the branch-wise variants. The Bilinear Lag-Lead (BT-LL) framework thus provides an explicit, frequency response mechanism for integrating contributions of low- and high-frequency information in pulmonary nodule representations.
- 22 September 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1934912
Open AccessFrequency response compensation to optimize pretrained deep learning models in lung nodule malignancy classification
Purpose: Accurate differentiation of benign and malignant pulmonary nodules on computed tomography (CT) is important for early lung-cancer assessment. However, classification remains challenging because benign and malignant nodules may exhibit subtle differences in texture, margin irregularity, boundary sharpness, and internal morphology that may not be fully represented by conventional convolutional feature extraction. Methods: A frequency response Bilinear Lag-Lead (BT-LL) EfficientNetB3 framework was developed for benign-malignant pulmonary nodule classification. The proposed approach integrates Lag-Lead frequency compensation into the pretrained feature-extraction process to refine intermediate representations. The Lag component performs structural smoothing to preserve low-frequency information, whereas the Lead component enhances high-frequency boundary and texture characteristics. These corresponding representations are integrated with the original feature maps through a dedicated fusion operation to obtain frequency-refined features for classification. The compensation filters are fixed, thereby avoiding additional parameter learning within the filtering operation. Patient-level data partitioning was used to ensure that nodules from the same patient were not distributed across different experimental subsets, reducing the risk of information leakage. Results: The proposed model achieved an accuracy, sensitivity, and specificity of 0.90, an F1-score of 0.81, and an AUROC of 0.95 on the held-out patient-level test set. Performance was additionally assessed using patient-level cross-validation, cross imbalance analysis, an IQ-OTH/NCCD image-level evaluation and an independent BUS-BRA patient-level evaluation. The experimental analysis further included statistical significance testing using McNemar’s test for pairwise model comparisons, followed by the Holm correction to control for multiple comparisons, while Cochran’s Q test was employed to assess performance differences across multiple model configurations. The combined configuration achieved the best sensitivity and balanced accuracy relative to the branch-wise variants. The Bilinear Lag-Lead (BT-LL) framework thus provides an explicit, frequency response mechanism for integrating contributions of low- and high-frequency information in pulmonary nodule representations.
- Open Access
- 24 June 2026
- Published by Elsevier in Intelligence-Based Medicine
- Vol. 15
Fractal fractional attention UNet based feature enhancement for accurate lung nodule segmentation
- 24 June 2026
- Published by Elsevier in Intelligence-Based Medicine
- Vol. 15
Open AccessFractal fractional attention UNet based feature enhancement for accurate lung nodule segmentation
- Open Access
- 30 April 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1794876
Feature stabilization in convolutional neural networks using Proportional Integral Controller for lung nodule classification
Introduction: Reliable classification of lung nodules from computed tomography (CT) images remains a challenging problem due to variations in image intensity, noise, and unstable feature representations during deep network training. Although convolutional neural networks (CNNs) have achieved promising results in medical image analysis, their internal feature dynamics are often difficult to control, which can affect convergence stability and generalization, particularly when working with limited clinical data. Method: In this work, we propose a control-inspired CNN framework that incorporates a Proportional Integral Controller (PIC) to regulate feature representations during the learning process. The PIC is integrated into the network in two distinct ways: as a preprocessing module before the CNN and as an intermediate layer embedded within the convolutional architecture. Both manually tuned and automatically learned PIC configurations are investigated to analyze the influence of fixed, knowledge-driven control parameters vs. adaptive, data-driven feedback mechanisms. The proportional component responds to instantaneous feature deviations, while the integral component compensates for accumulated errors, jointly contributing to more stable and consistent feature learning. Result: The proposed approach is evaluated on the IQ-OTH/NCCD lung cancer dataset using standard classification metrics. The proposed method achieves state-of-the-art performance (Accuracy 0.96, F1-score 0.96) and eliminates false positives (Precision 0.93). Ablation and statistical analyses confirm the importance of PIC placement and parameter tuning, while cross-dataset validation demonstrates strong generalization. Overall, this study demonstrates that integrating principles from control theory into deep learning architectures provides an effective and interpretable strategy for enhancing medical image classification.
- 30 April 2026
- Published by Frontiers Media SA in Frontiers in Artificial Intelligence
- Vol. 9, 1794876
Open AccessFeature stabilization in convolutional neural networks using Proportional Integral Controller for lung nodule classification
Introduction: Reliable classification of lung nodules from computed tomography (CT) images remains a challenging problem due to variations in image intensity, noise, and unstable feature representations during deep network training. Although convolutional neural networks (CNNs) have achieved promising results in medical image analysis, their internal feature dynamics are often difficult to control, which can affect convergence stability and generalization, particularly when working with limited clinical data. Method: In this work, we propose a control-inspired CNN framework that incorporates a Proportional Integral Controller (PIC) to regulate feature representations during the learning process. The PIC is integrated into the network in two distinct ways: as a preprocessing module before the CNN and as an intermediate layer embedded within the convolutional architecture. Both manually tuned and automatically learned PIC configurations are investigated to analyze the influence of fixed, knowledge-driven control parameters vs. adaptive, data-driven feedback mechanisms. The proportional component responds to instantaneous feature deviations, while the integral component compensates for accumulated errors, jointly contributing to more stable and consistent feature learning. Result: The proposed approach is evaluated on the IQ-OTH/NCCD lung cancer dataset using standard classification metrics. The proposed method achieves state-of-the-art performance (Accuracy 0.96, F1-score 0.96) and eliminates false positives (Precision 0.93). Ablation and statistical analyses confirm the importance of PIC placement and parameter tuning, while cross-dataset validation demonstrates strong generalization. Overall, this study demonstrates that integrating principles from control theory into deep learning architectures provides an effective and interpretable strategy for enhancing medical image classification.
- 1 October 2025
- Published by Springer Nature
- p. 441-452
Improving the Transmission Performance to Minimizing the Latency in Mobile ad hoc Network
- 1 October 2025
- Published by Springer Nature
- p. 441-452
Improving the Transmission Performance to Minimizing the Latency in Mobile ad hoc Network
- Open Access
- 26 May 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-30
Smart neural network and cognitive computing process for multi task nuclei detection segmentation and classification in breast cancer histopathology images
The detection, segmentation, and differentiation of benign and malignant nuclei from the histopathology images is a challenging task for the early diagnosis of breast cancer. Misinterpretation of True Negative (TN) and False Positive (FP) can generate incorrect results. The proposed Cognitive Computing Process (CCP) detects and segments the nuclei using Deep U-Net with Spatial Attention Mechanisms (SAM) and microns-per-pixel measurements to accurately locate and assess nuclei density. To separate the nuclei of benign and malignant, the patches are introduced to leverage the model’s learning process. The proposed Smart Neural Network (SNN) models contain Smart Convolutional Neural Network (SCNN) and Deep Convolutional Neural Network (DCNN) to reduce incorrect results. Proposed CCP and SNN were evaluated using the BreakHis dataset, which contains 5547 images of benign and malignant samples at various magnifications (40×, 100×, 200×, 400×). These images were processed into patches, totaling 11,642, 9282, 9102, and 9678 patches, each 224 × 224 pixels. The CCP model outperformed state-of-the-art models UNet, Residual UNet (ResUNet), and Convolutional Neural Network Long Short-Term Memory (CNN-LSTM) with a Dice coefficient of 99.90%, an F1-score of 99.04%, a precision of 99.80%, and a recall of 99.76%. The learning process began with a learning rate of 0.01 and a decay rate of 0.8, and the SCNN achieved false negative and false positive rates of 0.04 and 0.05 for low-density nuclei at 400× and 40× magnification, respectively. In contrast, the Deep Convolutional Neural Network (DCNN) recorded rates of 0.02 and 0.01. For high-density patches, the SCNN model FN and FP rates of 0.0 and 0.08, while the DCNN reported 0.09 and 0.0. The proposed learning process with Smart Neural Networks (SNN) achieved high precision (77–99%), recall (75–99%), F1-score (75–99%), and an AUC of 86–100%. The combination of CCP and SNN improved accuracy over existing CNN models like ResNet50, VGG19, DenseNet109, DenseNet201, and VGG16. An ablation study showed a p-value of 0.00003 based on the AUC, highlighting the model’s potential to enhance automated breast cancer diagnosis and support clinical decision-making.
- 26 May 2025
- Published by Springer Nature in Scientific Reports
- Vol. 15 (1) , 1-30
Open AccessSmart neural network and cognitive computing process for multi task nuclei detection segmentation and classification in breast cancer histopathology images
The detection, segmentation, and differentiation of benign and malignant nuclei from the histopathology images is a challenging task for the early diagnosis of breast cancer. Misinterpretation of True Negative (TN) and False Positive (FP) can generate incorrect results. The proposed Cognitive Computing Process (CCP) detects and segments the nuclei using Deep U-Net with Spatial Attention Mechanisms (SAM) and microns-per-pixel measurements to accurately locate and assess nuclei density. To separate the nuclei of benign and malignant, the patches are introduced to leverage the model’s learning process. The proposed Smart Neural Network (SNN) models contain Smart Convolutional Neural Network (SCNN) and Deep Convolutional Neural Network (DCNN) to reduce incorrect results. Proposed CCP and SNN were evaluated using the BreakHis dataset, which contains 5547 images of benign and malignant samples at various magnifications (40×, 100×, 200×, 400×). These images were processed into patches, totaling 11,642, 9282, 9102, and 9678 patches, each 224 × 224 pixels. The CCP model outperformed state-of-the-art models UNet, Residual UNet (ResUNet), and Convolutional Neural Network Long Short-Term Memory (CNN-LSTM) with a Dice coefficient of 99.90%, an F1-score of 99.04%, a precision of 99.80%, and a recall of 99.76%. The learning process began with a learning rate of 0.01 and a decay rate of 0.8, and the SCNN achieved false negative and false positive rates of 0.04 and 0.05 for low-density nuclei at 400× and 40× magnification, respectively. In contrast, the Deep Convolutional Neural Network (DCNN) recorded rates of 0.02 and 0.01. For high-density patches, the SCNN model FN and FP rates of 0.0 and 0.08, while the DCNN reported 0.09 and 0.0. The proposed learning process with Smart Neural Networks (SNN) achieved high precision (77–99%), recall (75–99%), F1-score (75–99%), and an AUC of 86–100%. The combination of CCP and SNN improved accuracy over existing CNN models like ResNet50, VGG19, DenseNet109, DenseNet201, and VGG16. An ablation study showed a p-value of 0.00003 based on the AUC, highlighting the model’s potential to enhance automated breast cancer diagnosis and support clinical decision-making.
- RetractionOpen Access
- 19 August 2024
- Published by Springer Nature in Soft Computing
- Vol. 28 (S1) , 239
Retraction Note: A heuristic method for initial dominant point detection for polygonal approximations
- 19 August 2024
- Published by Springer Nature in Soft Computing
- Vol. 28 (S1) , 239
RetractionOpen AccessRetraction Note: A heuristic method for initial dominant point detection for polygonal approximations
- 14 November 2023
- Published by Springer Nature
- p. 133-141
Geometric Transformations-Based Medical Image Augmentation
- 14 November 2023
- Published by Springer Nature
- p. 133-141
Geometric Transformations-Based Medical Image Augmentation