Her focus is on investigating the suitability of imaging techniques such as visible imaging, X-Ray imaging, HSI for conducting physical purity, and viability and vigor tests of different seeds. Her inquiry focuses on usability of the proposed image analysis based technique as an alternative to the e
9
Scopus Publications
203
Scholar Citations
3
Scholar h-index
3
Scholar i10-index
Scopus Publications
Secure and Efficient Federated Ensemble Learning for Hyperspectral Image Classification Anuj Kumar, Shveta Mahajan IETE Journal of Research, 2026 HSIs are widely employed in practical applications such as environmental monitoring, agriculture, military reconnaissance, and geographic mapping. Although Federated Learning (FL) allows decentralized model training without sharing raw data, transmitting model updates may still leak sensitive information. To address this issue, we proposed ADP-SSL (Adaptive Differentially Private Secure Socket Layer), a privacy-preserving FL framework that combines adaptive differential privacy with SSL-based communication. This design ensures secure transmission of model updates while reducing computational and communication overhead relative to traditional cryptographic techniques. Furthermore, to address the challenges posed by non-independent and identically distributed (non-IID) data across clients, we introduce FedEL (Federated Ensemble Learning). We evaluate the proposed approach on standard HSI datasets, including Indian Pines, Pavia University, and Barley. Experimental results show that ADP-SSL effectively mitigates privacy risks with minimal performance degradation, while FedEL improves classification accuracy in non-IID settings. Together, these methods offer a scalable and secure solution for real-world, distributed HSI classification.
Cracks Under the Lens of AI: A Deep Learning Review for Smarter Concrete Inspection Aditi Shah, Shveta Mahajan 2025 IEEE 2nd International Conference on Electrical Electronics Communication and Computers Elexcom 2025, 2025 Crack detection is a critical component of structural health monitoring for concrete infrastructure. Traditional inspection methods are often labor-intensive, subjective, and inefficient. The emergence of deep learning, particularly Convolutional Neural Networks, has introduced a revolutionary shift, enabling accurate, automated, and robust crack detection. This paper presents a comprehensive review of deep learning-based methods for concrete crack detection, synthesizing findings from recent literature. We examine three primary approaches: image classification for identifying crack presence, object detection for real-time localization using models like Faster R-CNN, YOLO and semantic segmentation for precise, pixel-level crack mapping with Fully Convolutional Networks. The pivotal role of transfer learning in overcoming data limitations is highlighted. Furthermore, we explore advanced applications such as the quantification of crack parameters like width and depth, and innovative methods for detecting internal defects. Key challenges including dataset scarcity, data imbalance, and model generalization under diverse environmental conditions are discussed. Finally, we outline promising future directions, including real-time integrated systems, unsupervised learning techniques, and the development of novel fused architectures for more efficient and comprehensive structural assessments.
Enhancement of Corn image quality using Very-Deep Super-Resolution (VDSR) neural network Harshit Rampal, Amitava Das, Shveta Mahajan, Satbir Singh Proceedings of the 3rd International Conference on Inventive Computation Technologies Icict 2018, 2018 Machine vision based quality inspection of food items is dependent on the quality of the acquired images. However, acquisition of high quality images requires costly and complex camera or other image acquisition equipment. In this paper, we have proposed a deep learning based solution which can be used to increase the resolution of input data acquired by low cost camera systems. The benefit of the proposed work lies in the fact that the input data acquired at lower resolution may be processed to obtain higher resolution suitable for detecting minute defects in agricultural products. Our proposed method was found to perform better than existing methods in terms of accuracy and visual appearance improvements.
A pre-processing based optimized edge weighting method for colour constancy Shveta Mahajan, Anu Rani, Mamta Sharma, Sudesh Kumar Mittal, Amitava Das Imaging Science Journal, 2018 An improvement in the existing weighted grey-edge (GE) framework for colour constancy is proposed. The acquired images are denoised by vector filtering and then, a two-step colour correction process is performed. In the first step, the GE method is used for estimating the global illuminant and perform the initial level of colour correction. The computed illuminant as well as the initial corrected image are used in the second step, which employs the weighted GE method to iteratively compute the final illuminant for obtaining the final colour corrected image. One hundred sixty-five standard test images from a publicly available colour constancy dataset were used to study the efficacy of the proposed algorithm. The results obtained indicate a significant improvement in the colour correction process as compared to the state-of-the-art colour constancy methods. The proposed algorithm reduced the mean angular error by approximately 67.85% compared to the existing weighted GE method.
Identification of poor visibility conditions in urban settings Hitesh Gudwani, Vikram Jit Singh, Shveta Mahajan, Deepti Mittal, Amitava Das 8th International Conference on Computing Communications and Networking Technologies Icccnt 2017, 2017 Street lighting provides a number of benefits. It can be used to promote security in urban areas and to increase the quality of life. Street light also provides safety for drivers, riders and pedestrians. In this paper, a machine vision system is proposed to identify the variations in the lighting of the road. The objective of this paper is to monitor the variations in the average intensity of the light on the road. The video data was acquired of night time road conditions and frames were extracted from video data. Several image processing techniques such as adaptive histogram equalization and morphological operations were performed on the images and the average intensity of the street light on each frame was computed. In this study, the variation in average intensity values are found to be useful in identifying the low lighting conditions on the road. The preliminary results clearly indicate the variations in the light falling on the road during night time.
Internal crack detection in kidney bean seeds using X-ray imaging technique Surbhi Sood, Shveta Mahajan, Amit Doegar, Amitava Das 2016 International Conference on Advances in Computing Communications and Informatics Icacci 2016, 2016 Seed quality testing is a contemporary research area that is motivated towards increasing agricultural productivity. For accurate quality assessment of seeds, internal morphological characteristics should be thoroughly examined in addition to the external examination. The soft X-ray imaging technique enables the visualization of the internal morphological attributes of agricultural seeds and grains in a non-destructive manner. The objective of this paper was to study the efficacy of using the X-ray imaging technique to detect the internal cracks in kidney bean seeds. The X-ray images of the sample seeds were acquired and image processing techniques such as histogram thresholding and morphological operations were applied on them. The segmented seed images were further processed and features were extracted. The extracted features were utilized for automatic detection of internal cracks, if present. The obtained results clearly indicated the usability of X-ray imaging techniques for automatic non-destructive detection of internal cracks in kidney bean seeds, as an essential component of their quality assessment.
An efficient directional distance filter for multi-band images Shveta Mahajan, Amitava Das 2nd International Conference on Signal Processing and Integrated Networks Spin 2015, 2015 An efficient directional distance filter (EDDF) for multi-band color images is introduced to remove Gaussian and impulse noise. The proposed filter employs each pixel color component simultaneously and maintains the tradeoff between the edge preservation and noise removal. The noise filtering technique is improved by selectively using the average, the vector median or the Gaussian filter based on minimal directional distance criteria. The principle behind the proposed filter is explained and comparison with the other popular vector filter is provided. The results are tested on images from publicly available data sets. The proposed filter performs better than directional distance filter (DDF) in terms of VRMSE. The obtained results indicate that the proposed EDDF can remove the impulse noise and Gaussian noise from RGB color images more effectively than the existing state-of-the-art technique.
Enhancement of Corn image quality using Very-Deep Super-Resolution (VDSR) neural network H Rampal, A Das, S Mahajan, S Singh 2018 3rd International Conference on Inventive Computation Technologies … , 2018 2018 Citations: 2
Machine vision based alternative testing approach for physical purity, viability and vigour testing of soybean seeds ( Glycine max ) S Mahajan, SK Mittal, A Das Journal of food science and technology 55 (10), 3949-3959 , 2018 2018 Citations: 39
A pre-processing based optimized edge weighting method for colour constancy S Mahajan, A Rani, M Sharma, SK Mittal, A Das The Imaging Science Journal 66 (4), 231-238 , 2018 2018 Citations: 1
Identification of poor visibility conditions in urban settings H Gudwani, VJ Singh, S Mahajan, D Mittal, A Das 2017 8th International Conference on Computing, Communication and Networking … , 2017 2017 Citations: 1
Internal crack detection in kidney bean seeds using X-ray imaging technique S Sood, S Mahajan, A Doegar, A Das 2016 International Conference on Advances in Computing, Communications and … , 2016 2016 Citations: 20
Image acquisition techniques for assessment of legume quality S Mahajan, A Das, HK Sardana Trends in Food Science & Technology 42 (2), 116-133 , 2015 2015 Citations: 140
An efficient directional distance filter for multi-band images S Mahajan, A Das 2015 2nd International Conference on Signal Processing and Integrated … , 2015 2015
MOST CITED SCHOLAR PUBLICATIONS
Image acquisition techniques for assessment of legume quality S Mahajan, A Das, HK Sardana Trends in Food Science & Technology 42 (2), 116-133 , 2015 2015 Citations: 140
Machine vision based alternative testing approach for physical purity, viability and vigour testing of soybean seeds ( Glycine max ) S Mahajan, SK Mittal, A Das Journal of food science and technology 55 (10), 3949-3959 , 2018 2018 Citations: 39
Internal crack detection in kidney bean seeds using X-ray imaging technique S Sood, S Mahajan, A Doegar, A Das 2016 International Conference on Advances in Computing, Communications and … , 2016 2016 Citations: 20
Enhancement of Corn image quality using Very-Deep Super-Resolution (VDSR) neural network H Rampal, A Das, S Mahajan, S Singh 2018 3rd International Conference on Inventive Computation Technologies … , 2018 2018 Citations: 2
A pre-processing based optimized edge weighting method for colour constancy S Mahajan, A Rani, M Sharma, SK Mittal, A Das The Imaging Science Journal 66 (4), 231-238 , 2018 2018 Citations: 1
Identification of poor visibility conditions in urban settings H Gudwani, VJ Singh, S Mahajan, D Mittal, A Das 2017 8th International Conference on Computing, Communication and Networking … , 2017 2017 Citations: 1
An efficient directional distance filter for multi-band images S Mahajan, A Das 2015 2nd International Conference on Signal Processing and Integrated … , 2015 2015
