Sushma Jain
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sushma Jain has a total of 29 co-authors. Here are the top co-authors by number of shared publications:
- Shivani Goel
School of CS and AI, SR University
3 shared publications - Amandeep Kaur
National Institute of Technology Kurukshetra, Kurukshetra, India
2 shared publications - Amandeep Singh Bhatia
Chitkara University Institute of Engineering & Technology Chitkara University Punjab India
2 shared publications - Archana Singh
Department of Computer Science & Engineering, Thapar Institute of Engineering & Technology, Patiala 147004, India
2 shared publications - Rakesh Sharda
Department of Soil and Water Engineering, Punjab Agricultural University, Ludhiana, India
2 shared publications - Sukhchandan Randhawa
Computer Science and Engineering Department, Thapar Institute of Engineering and Technology, Patiala, India
2 shared publications
Latest Publications
- 2 January 2026
- Published by Springer Nature
- p. 183-195
Enhancing Tomato Leaf Disease Classification with Deep Transfer Learning on VGG16
- 2 January 2026
- Published by Springer Nature
- p. 183-195
Enhancing Tomato Leaf Disease Classification with Deep Transfer Learning on VGG16
- Open Access
- 17 October 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (20) , 2619
Pixel-Attention W-Shaped Network for Joint Lesion Segmentation and Diabetic Retinopathy Severity Staging
Background: Visual impairment remains a critical public health challenge, and diabetic retinopathy (DR) is a leading cause of preventable blindness worldwide. Early stages of the disease are particularly difficult to identify, as lesions are subtle, expert review is time-consuming, and conventional diagnostic workflows remain subjective. Methods: To address these challenges, we propose a novel Pixel-Attention W-shaped (PAW-Net) deep learning framework that integrates a Lesion-Prior Cross Attention (LPCA) module with a W-shaped encoder–decoder architecture. The LPCA module enhances pixel-level representation of microaneurysms, hemorrhages, and exudates, while the dual-branch W-shaped design jointly performs lesion segmentation and disease severity grading in a single, clinically interpretable pass. The framework has been trained and validated using DDR and a preprocessed Messidor + EyePACS dataset, with APTOS-2019 reserved for external, out-of-distribution evaluation. Results: The proposed PAW-Net framework achieved robust performance across severity levels, with an accuracy of 98.65%, precision of 98.42%, recall (sensitivity) of 98.83%, specificity of 99.12%, F1-score of 98.61%, and a Dice coefficient of 98.61%. Comparative analyses demonstrate consistent improvements over contemporary architectures, particularly in accuracy and F1-score. Conclusions: The PAW-Net framework generates interpretable lesion overlays that facilitate rapid triage and follow-up, exhibits resilience under domain shift, and maintains an efficient computational footprint suitable for telemedicine and mobile deployment.
- 17 October 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (20) , 2619
Open AccessPixel-Attention W-Shaped Network for Joint Lesion Segmentation and Diabetic Retinopathy Severity Staging
Background: Visual impairment remains a critical public health challenge, and diabetic retinopathy (DR) is a leading cause of preventable blindness worldwide. Early stages of the disease are particularly difficult to identify, as lesions are subtle, expert review is time-consuming, and conventional diagnostic workflows remain subjective. Methods: To address these challenges, we propose a novel Pixel-Attention W-shaped (PAW-Net) deep learning framework that integrates a Lesion-Prior Cross Attention (LPCA) module with a W-shaped encoder–decoder architecture. The LPCA module enhances pixel-level representation of microaneurysms, hemorrhages, and exudates, while the dual-branch W-shaped design jointly performs lesion segmentation and disease severity grading in a single, clinically interpretable pass. The framework has been trained and validated using DDR and a preprocessed Messidor + EyePACS dataset, with APTOS-2019 reserved for external, out-of-distribution evaluation. Results: The proposed PAW-Net framework achieved robust performance across severity levels, with an accuracy of 98.65%, precision of 98.42%, recall (sensitivity) of 98.83%, specificity of 99.12%, F1-score of 98.61%, and a Dice coefficient of 98.61%. Comparative analyses demonstrate consistent improvements over contemporary architectures, particularly in accuracy and F1-score. Conclusions: The PAW-Net framework generates interpretable lesion overlays that facilitate rapid triage and follow-up, exhibits resilience under domain shift, and maintains an efficient computational footprint suitable for telemedicine and mobile deployment.
- Open Access
- 17 September 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (18) , 2355
A Multi-Model Image Enhancement and Tailored U-Net Architecture for Robust Diabetic Retinopathy Grading
Background: Diabetic retinopathy (DR) is a leading cause of preventable vision impairment in individuals with diabetes. Early detection is essential, yet often hindered by subtle disease progression and reliance on manual expert screening. This study introduces an AI-based framework designed to achieve robust multiclass DR classification from retinal fundus images, addressing the challenges of early diagnosis and fine-grained lesion discrimination. Methods: The framework incorporates preprocessing steps such as pixel intensity normalization and geometric correction. A Hybrid Local-Global Retina Super-Resolution (HLG-RetinaSR) module is developed, combining deformable convolutional networks for local lesion enhancement with vision transformers for global contextual representation. Classification is performed using a hierarchical approach that integrates three models: a Convolutional Neural Network (CNN), DenseNet-121, and a custom multi-branch RefineNet-U architecture. Results: Experimental evaluation demonstrates that the combined HLG-RetinaSR and RefineNet-U approach consistently achieves precision, recall, F1-score, and accuracy values exceeding 99% across all DR severity levels. The system effectively emphasizes vascular abnormalities while suppressing background noise, surpassing existing state-of-the-art methods in accuracy and robustness. Conclusions: The proposed hybrid pipeline delivers a scalable, interpretable, and clinically relevant solution for DR screening. By improving diagnostic reliability and supporting early intervention, the system holds strong potential to assist ophthalmologists in reducing preventable vision loss.
- 17 September 2025
- Published by MDPI AG in Diagnostics
- Vol. 15 (18) , 2355
Open AccessA Multi-Model Image Enhancement and Tailored U-Net Architecture for Robust Diabetic Retinopathy Grading
Background: Diabetic retinopathy (DR) is a leading cause of preventable vision impairment in individuals with diabetes. Early detection is essential, yet often hindered by subtle disease progression and reliance on manual expert screening. This study introduces an AI-based framework designed to achieve robust multiclass DR classification from retinal fundus images, addressing the challenges of early diagnosis and fine-grained lesion discrimination. Methods: The framework incorporates preprocessing steps such as pixel intensity normalization and geometric correction. A Hybrid Local-Global Retina Super-Resolution (HLG-RetinaSR) module is developed, combining deformable convolutional networks for local lesion enhancement with vision transformers for global contextual representation. Classification is performed using a hierarchical approach that integrates three models: a Convolutional Neural Network (CNN), DenseNet-121, and a custom multi-branch RefineNet-U architecture. Results: Experimental evaluation demonstrates that the combined HLG-RetinaSR and RefineNet-U approach consistently achieves precision, recall, F1-score, and accuracy values exceeding 99% across all DR severity levels. The system effectively emphasizes vascular abnormalities while suppressing background noise, surpassing existing state-of-the-art methods in accuracy and robustness. Conclusions: The proposed hybrid pipeline delivers a scalable, interpretable, and clinically relevant solution for DR screening. By improving diagnostic reliability and supporting early intervention, the system holds strong potential to assist ophthalmologists in reducing preventable vision loss.
- 30 October 2023
- Published by World Scientific Pub Co Pte Ltd in Journal of Mechanics in Medicine and Biology
- Vol. 24 (06)
PREDICTING CARDIAC HEALTH USING SUB-COMPONENT OF A PHONOCARDIOGRAM
There has been a steady rise in the number of deaths throughout the world due to heart diseases. This can be mitigated, to a large extent, if cardiovascular disorders can be detected timely and efficiently. Electrocardiograms (ECGs) and phonocardiograms (PCGs) are the two most popular diagnostic tools used for detecting cardiac problems. Another simple and efficient method for quickly identifying cardiovascular illness is Auscultation. In this work, the cardiac sound signal has been transformed into its equivalent spectrogram representation for detecting cardiac problems. The novelty of the proposed approach is the deployment of customized transfer learning (TL) models on sub-component of a spectrogram called Harmonic Spectrogram, instead of taking full spectrogram. Experiments have been conducted using PhysioNet 2016, which is considered a benchmark dataset. TL models, viz. MobileNet, DenseNet121, InceptionResnetV2, VGG16, and InceptionV3 have been put to use for categorizing cardiac sound waves as normal or pathological. The results exhibit that the MobileNet has achieved greater accuracy (93.45%), recall (92.46%), Precision (97.82%), F1 Score (95.06%) than many of the peers.
- 30 October 2023
- Published by World Scientific Pub Co Pte Ltd in Journal of Mechanics in Medicine and Biology
- Vol. 24 (06)
PREDICTING CARDIAC HEALTH USING SUB-COMPONENT OF A PHONOCARDIOGRAM
There has been a steady rise in the number of deaths throughout the world due to heart diseases. This can be mitigated, to a large extent, if cardiovascular disorders can be detected timely and efficiently. Electrocardiograms (ECGs) and phonocardiograms (PCGs) are the two most popular diagnostic tools used for detecting cardiac problems. Another simple and efficient method for quickly identifying cardiovascular illness is Auscultation. In this work, the cardiac sound signal has been transformed into its equivalent spectrogram representation for detecting cardiac problems. The novelty of the proposed approach is the deployment of customized transfer learning (TL) models on sub-component of a spectrogram called Harmonic Spectrogram, instead of taking full spectrogram. Experiments have been conducted using PhysioNet 2016, which is considered a benchmark dataset. TL models, viz. MobileNet, DenseNet121, InceptionResnetV2, VGG16, and InceptionV3 have been put to use for categorizing cardiac sound waves as normal or pathological. The results exhibit that the MobileNet has achieved greater accuracy (93.45%), recall (92.46%), Precision (97.82%), F1 Score (95.06%) than many of the peers.
- Open Access
- 30 August 2022
- Published by Blue Eyes Intelligence Engineering and Sciences Engineering and Sciences Publication (BEIESP) in International Journal of Engineering and Advanced Technology
- Vol. 11 (6) , 168-173
Assessing the Influence of the Covid 19 Pandemic on Indian Pharmaceutical Companies
Our objective is to study, analyze and draw inferences on the movement of the stock prices of Indian pharmaceutical companies solely based on the COVID-19 pandemic in India. We specifically targeted pharmaceutical stocks because their share price is more directly dependent on the COVID-19 pandemic than companies in other sectors. As the demand for the primary products sold by pharmaceutical companies, i.e., medicines, is directly dependent on the COVID-19 pandemic, a common hypothesis is that the stock prices of pharmaceutical companies at a given time are significantly contingent upon the COVID-19 pandemic situation. We have tested this hypothesis by calculating the correlation between pharmaceutical stock prices and COVID-19 variables that measure the severity and provide an outline of the COVID-19 pandemic. The COVID-19 variables we have considered provide information regarding covid cases, deaths, testing, vaccination, positivity rate, virus reproduction rate, and government restrictions to counter the spread of the virus. Furthermore, as human emotion plays a significant part in deciding the share prices, we have considered public fear and awareness by considering the frequency by which the terms “Covid 19” and “Covid medicines” are searched on Google. We have considered the stock prices of 19 companies that contribute to the Nifty Pharma index as the target values upon which the impact of the COVID-19 pandemic is tested. We have selected the covid fields that have the most significant impact on the pharmaceutical stock prices and then calculated the correlation between the Covid fields and the stock parameters. The data we have considered for our study belongs to the period from 15th March 2020 to 17th February 2022. Link to the Github repository with the code for the presented research: https://github.com/Atishaysjain/Predictive-Analysis-Project.
- 30 August 2022
- Published by Blue Eyes Intelligence Engineering and Sciences Engineering and Sciences Publication (BEIESP) in International Journal of Engineering and Advanced Technology
- Vol. 11 (6) , 168-173
Open AccessAssessing the Influence of the Covid 19 Pandemic on Indian Pharmaceutical Companies
Our objective is to study, analyze and draw inferences on the movement of the stock prices of Indian pharmaceutical companies solely based on the COVID-19 pandemic in India. We specifically targeted pharmaceutical stocks because their share price is more directly dependent on the COVID-19 pandemic than companies in other sectors. As the demand for the primary products sold by pharmaceutical companies, i.e., medicines, is directly dependent on the COVID-19 pandemic, a common hypothesis is that the stock prices of pharmaceutical companies at a given time are significantly contingent upon the COVID-19 pandemic situation. We have tested this hypothesis by calculating the correlation between pharmaceutical stock prices and COVID-19 variables that measure the severity and provide an outline of the COVID-19 pandemic. The COVID-19 variables we have considered provide information regarding covid cases, deaths, testing, vaccination, positivity rate, virus reproduction rate, and government restrictions to counter the spread of the virus. Furthermore, as human emotion plays a significant part in deciding the share prices, we have considered public fear and awareness by considering the frequency by which the terms “Covid 19” and “Covid medicines” are searched on Google. We have considered the stock prices of 19 companies that contribute to the Nifty Pharma index as the target values upon which the impact of the COVID-19 pandemic is tested. We have selected the covid fields that have the most significant impact on the pharmaceutical stock prices and then calculated the correlation between the Covid fields and the stock parameters. The data we have considered for our study belongs to the period from 15th March 2020 to 17th February 2022. Link to the Github repository with the code for the presented research: https://github.com/Atishaysjain/Predictive-Analysis-Project.
- 20 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 13, 621-626
Predicting Evapotranspiration (ETo) by Training Regression Models
Evapotranspiration is a critical natural process in agriculture. Every farmer has to deal with this in every stage of crop development. Proper estimation of soil and leaf’s-based Evapotranspiration (ETo) is required to be done. Traditionally it is being done based on climate forecasting only, but nowadays, many new techniques have been shown up by which this can be estimated more accurately. In this study, machine learning and deep learning models are being used to determine the ETo. Regression techniques are preferred over classification so that every possible value can be taken as an input for accuracy. Nowadays, many algorithms have proved their efficiency with the specific dataset. There is no unique algorithm that has established itself as suitable for every type of dataset. This study will help to find the best algorithm for getting accurate predictions on ETo. A few famous regression algorithms are being opted for this study, and the best resultant algorithm is determined for a given dataset. The highlighted algorithm based on the experiment should only be followed if similar climate attributes are encountered for any research and study.
- 20 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- Vol. 13, 621-626
Predicting Evapotranspiration (ETo) by Training Regression Models
Evapotranspiration is a critical natural process in agriculture. Every farmer has to deal with this in every stage of crop development. Proper estimation of soil and leaf’s-based Evapotranspiration (ETo) is required to be done. Traditionally it is being done based on climate forecasting only, but nowadays, many new techniques have been shown up by which this can be estimated more accurately. In this study, machine learning and deep learning models are being used to determine the ETo. Regression techniques are preferred over classification so that every possible value can be taken as an input for accuracy. Nowadays, many algorithms have proved their efficiency with the specific dataset. There is no unique algorithm that has established itself as suitable for every type of dataset. This study will help to find the best algorithm for getting accurate predictions on ETo. A few famous regression algorithms are being opted for this study, and the best resultant algorithm is determined for a given dataset. The highlighted algorithm based on the experiment should only be followed if similar climate attributes are encountered for any research and study.
- 25 January 2022
- Published by Springer Nature in Journal of Ambient Intelligence and Humanized Computing
- Vol. 14 (7) , 8881-8897
Proposition of new ensemble data-intelligence model for evapotranspiration process simulation
- 25 January 2022
- Published by Springer Nature in Journal of Ambient Intelligence and Humanized Computing
- Vol. 14 (7) , 8881-8897
Proposition of new ensemble data-intelligence model for evapotranspiration process simulation
- 16 April 2021
- Published by Bentham Science Publishers Ltd. in Recent Advances in Electrical & Electronic Engineering (Formerly Recent Patents on Electrical & Electronic Engineering)
- Vol. 14 (3) , 290-303
A Review on Machine-learning Based Code Smell Detection Techniques in Object-oriented Software System(s)
Background: Code smells are symptoms that something may be wrong in software systems that can cause complications in maintaining software quality. In literature, there exist many code smells and their identification is far from trivial. Thus, several techniques have also been proposed to automate code smell detection in order to improve software quality. Objective: This paper presents an up-to-date review of simple and hybrid machine learning-based code smell detection techniques and tools. Methods: We collected all the relevant research published in this field till 2020. We extracted the data from those articles and classified them into two major categories. In addition, we compared the selected studies based on several aspects like code smells, machine learning techniques, datasets, programming languages used by datasets, dataset size, evaluation approach, and statistical testing. Results: A majority of empirical studies have proposed machine-learning based code smell detection tools. Support vector machine and decision tree algorithms are frequently used by the researchers. Along with this, a major proportion of research is conducted on Open Source Softwares (OSS) such as Xerces, Gantt Project and ArgoUml. Furthermore, researchers pay more attention to Feature Envy and Long Method code smells. Conclusion: We identified several areas of open research like the need for code smell detection techniques using hybrid approaches, the need for employing valid industrial datasets, etc.
- 16 April 2021
- Published by Bentham Science Publishers Ltd. in Recent Advances in Electrical & Electronic Engineering (Formerly Recent Patents on Electrical & Electronic Engineering)
- Vol. 14 (3) , 290-303
A Review on Machine-learning Based Code Smell Detection Techniques in Object-oriented Software System(s)
Background: Code smells are symptoms that something may be wrong in software systems that can cause complications in maintaining software quality. In literature, there exist many code smells and their identification is far from trivial. Thus, several techniques have also been proposed to automate code smell detection in order to improve software quality. Objective: This paper presents an up-to-date review of simple and hybrid machine learning-based code smell detection techniques and tools. Methods: We collected all the relevant research published in this field till 2020. We extracted the data from those articles and classified them into two major categories. In addition, we compared the selected studies based on several aspects like code smells, machine learning techniques, datasets, programming languages used by datasets, dataset size, evaluation approach, and statistical testing. Results: A majority of empirical studies have proposed machine-learning based code smell detection tools. Support vector machine and decision tree algorithms are frequently used by the researchers. Along with this, a major proportion of research is conducted on Open Source Softwares (OSS) such as Xerces, Gantt Project and ArgoUml. Furthermore, researchers pay more attention to Feature Envy and Long Method code smells. Conclusion: We identified several areas of open research like the need for code smell detection techniques using hybrid approaches, the need for employing valid industrial datasets, etc.
- 15 August 2020
- Published by Wiley in Transactions on Emerging Telecommunications Technologies
- Vol. 32 (6)
SecRIP: Secure and reliable intercluster routing protocol for efficient data transmission in flying ad hoc networks
Unmanned aerial vehicles (UAVs) come up with eminent opportunities for military and civilian applications. Such opportunities come with unique challenges. The high mobility of UAVs leads to the frequent change in the network topology that results in packet loss or routing path failure, and so on. Moreover, secure routing is one of the most important features while organizing wireless communication among UAVs. To counter these challenges, a secure and reliable routing protocol (SecRIP) for the flying ad hoc network is proposed for efficient and reliable data transfer. This SecRIP works toward the enhancement of the quality of service (QoS) and quality of experience (QoE) metrics. The SecRIP works on two algorithms: (i) a chaotic algae algorithm and (ii) dragonfly algorithm; these algorithm serves the functionality of cluster selection, management, and data transmission in intercluster. The proposed technique is compared with other existing techniques using a network simulator. The results show that the proposed technique is capable of maintaining the higher QoS and the QoE standards and also helps the nodes conserve their power without compromising on the performance. The simulated results show improvement as it decreases the delay incurred in SecRIP is 24%, and routing overhead is 19%, where the PDR of SecRIP is increased by 32% in comparison with existing protocols. In an extreme case, when node density increases, the SecRIP helps to limits the delay to 41%, and the PDR is 28% higher than the other existing protocols. Finally, the result demonstrates that the SecRIP protocol achieves secure and reliable data transmission.
- 15 August 2020
- Published by Wiley in Transactions on Emerging Telecommunications Technologies
- Vol. 32 (6)
SecRIP: Secure and reliable intercluster routing protocol for efficient data transmission in flying ad hoc networks
Unmanned aerial vehicles (UAVs) come up with eminent opportunities for military and civilian applications. Such opportunities come with unique challenges. The high mobility of UAVs leads to the frequent change in the network topology that results in packet loss or routing path failure, and so on. Moreover, secure routing is one of the most important features while organizing wireless communication among UAVs. To counter these challenges, a secure and reliable routing protocol (SecRIP) for the flying ad hoc network is proposed for efficient and reliable data transfer. This SecRIP works toward the enhancement of the quality of service (QoS) and quality of experience (QoE) metrics. The SecRIP works on two algorithms: (i) a chaotic algae algorithm and (ii) dragonfly algorithm; these algorithm serves the functionality of cluster selection, management, and data transmission in intercluster. The proposed technique is compared with other existing techniques using a network simulator. The results show that the proposed technique is capable of maintaining the higher QoS and the QoE standards and also helps the nodes conserve their power without compromising on the performance. The simulated results show improvement as it decreases the delay incurred in SecRIP is 24%, and routing overhead is 19%, where the PDR of SecRIP is increased by 32% in comparison with existing protocols. In an extreme case, when node density increases, the SecRIP helps to limits the delay to 41%, and the PDR is 28% higher than the other existing protocols. Finally, the result demonstrates that the SecRIP protocol achieves secure and reliable data transmission.
- 29 May 2020
- Published by Springer Nature in The International Journal of Advanced Manufacturing Technology
- Vol. 108 (5-6) , 1563-1585
Hybrid frameworks for flexible job shop scheduling
- 29 May 2020
- Published by Springer Nature in The International Journal of Advanced Manufacturing Technology
- Vol. 108 (5-6) , 1563-1585
Hybrid frameworks for flexible job shop scheduling