Rajesh Kumar

EAAI Journal 2026 Journal Article

Spiking neural network-based energy-efficient framework for real-time robotic arm manipulation

• Ashok Kumar Saini
• Naveen Gehlot
• Rajesh Kumar
• Surender Hans
• Santosh Chaudhary
• Gulshan Sharma

In industrial applications, robotic arms are widely utilized, posing an important challenge in achieving precise positioning. Traditional solutions, such as inverse kinematics and polynomial trajectory generation, prove computationally costly and time-consuming for real-time control applications. This work proposes a bio-inspired spiking neural network (SNN) for controlling a 3-degree-of-freedom robotic arm without explicit pre-planning. Simulating spiking neurons using the leaky integrate-and-fire model balances biological realism with computational efficiency. The work includes tests with extreme targets, random coordinates, and practical experiments with robot arm hardware to cover the working area thoroughly. Numerical experiments confirm the efficiency of the proposed SNN, which solves the inverse kinematics in just 1. 50 ms, compared to 20. 15 ms for artificial neural network (ANN), when executed on the computational hardware platform with an Intel Core i7 central processing unit (2. 10 Gigahertz) and 16 Gigabytes of Random Access Memory. The reduction in time highlights SNN’s potential to optimize computational complexity, which further enhances the overall performance of robotic arm manipulators. Furthermore, across 3rd, 5th, and 7th order trajectories, the SNN consistently results in Mean Squared Error (MSE) for trajectory generation (0. 0046–0. 0900) and for inverse kinematics (0. 3245–1. 2474) as compared to ANN (0. 5990–2. 2320) and (1. 5015–2. 4870), respectively, confirming improved performance in both trajectory planning and inverse kinematics.

AIIM Journal 2025 Journal Article

Privacy-preserving federated transfer learning for enhanced liver lesion segmentation in PET–CT imaging

• Rajesh Kumar
• Shaoning Zeng
• Jay Kumar
• Zakria
• Xinfeng Mao

Positron Emission Tomography-Computed Tomography (PET–CT) evolution is critical for liver lesion diagnosis. However, data scarcity, privacy concerns, and cross-institutional imaging heterogeneity impede accurate deep learning model deployment. We propose a Federated Transfer Learning (FTL) framework that integrates federated learning’s privacy-preserving collaboration with transfer learning’s pre-trained model adaptation, enhancing liver lesion segmentation in PET–CT imaging. By leveraging a Feature Co-learning Block (FCB) and privacy-enhancing technologies (DP, HE), our approach ensures robust segmentation without sharing sensitive patient data. (1) A privacy-preserving FTL framework combining federated learning and adaptive transfer learning; (2) A multi-modal FCB for improved PET–CT feature integration; (3) Extensive evaluation across diverse institutions with privacy-enhancing technologies like Differential Privacy (DP) and Homomorphic Encryption (HE). Experiments on simulated multi-institutional PET–CT datasets demonstrate superior performance compared to baselines, with robust privacy guarantees. The FTL framework reduces data requirements and enhances generalizability, advancing liver lesion diagnostics.

EAAI Journal 2024 Journal Article

Artificial intelligence for computer aided detection of pneumoconiosis: A succinct review since 1974

• Faisel Mushtaq
• Saunak Bhattacharjee
• Sandeep Mandia
• Kuldeep Singh
• Satyendra Singh Chouhan
• Rajesh Kumar
• Priyanka Harjule

With the increasing availability of medical imaging data and advancements in artificial intelligence (AI) and computer vision (CV) techniques, computer aided diagnostic (CAD) systems have been consistently developed to help radiologists in the detection of pneumoconiosis. Pneumoconiosis is a respiratory disease caused by long term exposure of industrial dust. Pneumoconiosis has remained prevalent, even in countries with sophisticated healthcare systems and strict workplace measures. Consequently, this review article focuses on cascading the literature of existing CAD systems for pneumoconiosis diagnosis since 1974, providing a baseline for future research in this domain. For this purpose, 58 relevant articles were first selected after employing strict inclusion criteria, through 10 reliable databases and search engines, including Scopus, IEEE, Google Scholar, and PubMed etc. This review then systematically categorizes the selected CAD studies into two patterns, based on the employed methodology for pneumoconiosis diagnosis: texture-based and non-texture-based CAD systems. This study reveals that texture-based methods have been extensively adopted for pneumoconiosis diagnosis compared to non-texture-based methods (or deep learning based). However, deep learning approaches have shown superior performance thanks to the recent availability of large annotated CXR datasets and development of deep convolutional neural network (CNN) and transformer-based architectures. Finally, the analysis is concluded with a discussion on the shortcomings of current CAD systems and some suggested future directions for the development of effective diagnostic systems. Additionally, a number of benchmark datasets have also been discussed.

TIST Journal 2024 Journal Article

Responsible Recommendation Services with Blockchain Empowered Asynchronous Federated Learning

• Waqar Ali
• Rajesh Kumar
• Xiangmin Zhou
• Jie Shao

Privacy and trust are highly demanding in practical recommendation engines. Although Federated Learning (FL) has significantly addressed privacy concerns, commercial operators are still worried about several technical challenges while bringing FL into production. In addition, classical FL has several intrinsic operational limitations such as single-point failure, data and model tampering, and heterogenic clients participating in the FL process. To address these challenges in practical recommenders, we propose a responsible recommendation generation framework based on blockchain-empowered asynchronous FL that can be adopted for any model-based recommender system. In standard FL settings, we build an additional aggregation layer in which multiple trusted nodes guided by a mediator component perform gradient aggregation to achieve an optimal model locally in a parallel fashion. The mediator partitions users into K clusters, and each cluster is represented by a cluster head. Once a cluster gets semi-global convergence, the cluster head transmits model gradients to the FL server for global aggregation. In addition the trusted cluster heads are responsible to submit the converged semi-global model to a blockchain to ensure tamper resilience. In our settings, an additional mediator component works like an independent observer that monitors the performance of each cluster head, updates a reward score, and records it into a digital ledger. Finally, evaluation results on three diversified benchmarks illustrate that the recommendation performance on selected measures is considerably comparable with the standard and federated version of a well-known neural collaborative filtering recommender.

EAAI Journal 2024 Journal Article

Surface electromyography based explainable Artificial Intelligence fusion framework for feature selection of hand gesture recognition

• Naveen Gehlot
• Ashutosh Jena
• Ankit Vijayvargiya
• Rajesh Kumar

Over the past decade, the utilization of machine learning (ML) models for recognizing hand gestures from surface electromyography (sEMG) signals has been in demand for the control of prosthetics. Such real-time control demands swift responses and efficient models. The model’s efficiency heavily relies on selecting optimal handcrafted features. As the number of acquisition channels increases, the complexity of handcrafted features escalates the computational burden and necessitates the reduction of irrelevant or noisy features to enhance model efficiency. This study proposes an eXplainable Artificial Intelligence (XAI) fusion-based feature selection framework for sEMG-based hand gesture recognition (HGR). The proposed framework comprises two stages. Firstly, it combines three feature selection methods: filter, wrapper, and embedded. Secondly, it employs SHAP-based explainability for ML models to select relevant features. The first stage of feature selection retains ten relevant features, followed by a fine selection of 50% of features in the second stage. The performance of the proposed framework is compared with three baseline models. It shows that the proposed model performs better than other baseline models in terms of accuracy, precision, recall, f-score, and computational time. The proposed XAI fusion framework achieves an average classification accuracy of 80. 75% with the Extra Tree classifier, with a computational time of 3. 47 ms. Furthermore, to assess its robustness against baseline models, evaluation is conducted using a publicly available dataset, revealing superior performance compared to other baselines.

YNICL Journal 2018 Journal Article

Regional cortical thickness changes accompanying generalized tonic-clonic seizures

• Jennifer A. Ogren
• Raghav Tripathi
• Paul M. Macey
• Rajesh Kumar
• John M. Stern
• Dawn S. Eliashiv
• Luke A. Allen
• Beate Diehl

Objective: Generalized tonic-clonic seizures are accompanied by cardiovascular and respiratory sequelae that threaten survival. The frequency of these seizures is a major risk factor for sudden unexpected death in epilepsy (SUDEP), a leading cause of untimely death in epilepsy. The circumstances accompanying such fatal events suggest a cardiovascular or respiratory failure induced by unknown neural processes rather than an inherent cardiac or lung deficiency. Certain cortical regions, especially the insular, cingulate, and orbitofrontal cortices, are key structures that integrate sensory input and influence diencephalic and brainstem regions regulating blood pressure, cardiac rhythm, and respiration; output from those cortical regions compromised by epilepsy-associated injury may lead to cardiorespiratory dysregulation. The aim here was to assess changes in cortical integrity, reflected as cortical thickness, relative to healthy controls. Cortical alterations in areas that influence cardiorespiratory action could contribute to SUDEP mechanisms. Methods: < 0.05 corrected). Results: Increased cortical thickness appeared in post-central gyri, insula, and subgenual, anterior, posterior, and isthmus cingulate cortices. Post-central gyri increases were greater in females, while males showed more extensive cingulate increases. Frontal and temporal cortex, lateral orbitofrontal, frontal pole, and lateral parietal and occipital cortices showed thinning. The extents of thickness changes were sex- and hemisphere-dependent, with only males exhibiting right-sided and posterior cingulate thickening, while females showed only left lateral orbitofrontal thinning. Regional cortical thickness showed modest correlations with seizure frequency, but not epilepsy duration. Significance: Cortical thickening and thinning occur in patients with generalized tonic-clonic seizures, in cardiovascular and somatosensory areas, with extent of changes sex- and hemisphere-dependent. The data show injury in key autonomic and respiratory cortical areas, which may contribute to dysfunctional cardiorespiratory patterns during seizures, as well as to longer-term SUDEP risk.

YNICL Journal 2018 Journal Article

Sex-specific hippocampus volume changes in obstructive sleep apnea

• Paul M. Macey
• Janani P. Prasad
• Jennifer A. Ogren
• Ammar S. Moiyadi
• Ravi S. Aysola
• Rajesh Kumar
• Frisca L. Yan-Go
• Mary A. Woo

Introduction: Obstructive sleep apnea (OSA) patients show hippocampal-related autonomic and neurological symptoms, including impaired memory and depression, which differ by sex, and are mediated in distinct hippocampal subfields. Determining sites and extent of hippocampal sub-regional injury in OSA could reveal localized structural damage linked with OSA symptoms. Methods: High-resolution T1-weighted images were collected from 66 newly-diagnosed, untreated OSA (mean age ± SD: 46.3 ± 8.8 years; mean AHI ± SD: 34.1 ± 21.5 events/h;50 male) and 59 healthy age-matched control (46.8 ± 9.0 years;38 male) participants. We added age-matched controls with T1-weighted scans from two datasets (IXI, OASIS-MRI), for 979 controls total (426 male/46.5 ± 9.9 years). We segmented the hippocampus and analyzed surface structure with "FSL FIRST" software, scaling volumes for brain size, and evaluated group differences with ANCOVA (covariates: total-intracranial-volume, sex; P < .05, corrected). Results: In OSA relative to controls, the hippocampus showed small areas larger volume bilaterally in CA1 (surface displacement ≤0.56 mm), subiculum, and uncus, and smaller volume in right posterior CA3/dentate (≥ - 0.23 mm). OSA vs. control males showed higher bilateral volume (≤0.61 mm) throughout CA1 and subiculum, extending to head and tail, with greater right-sided increases; lower bilateral volumes (≥ - 0.45 mm) appeared in mid- and posterior-CA3/dentate. OSA vs control females showed only right-sided effects, with increased CA1 and subiculum/uncus volumes (≤0.67 mm), and decreased posterior CA3/dentate volumes (≥ - 0.52 mm). Unlike males, OSA females showed volume decreases in the right hippocampus head and tail. Conclusions: The hippocampus shows lateralized and sex-specific, OSA-related regional volume differences, which may contribute to sex-related expression of symptoms in the sleep disorder. Volume increases suggest inflammation and glial activation, whereas volume decreases suggest long-lasting neuronal injury; both processes may contribute to dysfunction in OSA.

YNICL Journal 2014 Journal Article

Brain putamen volume changes in newly-diagnosed patients with obstructive sleep apnea

• Rajesh Kumar
• Salar Farahvar
• Jennifer A. Ogren
• Paul M. Macey
• Paul M. Thompson
• Mary A. Woo
• Frisca L. Yan-Go
• Ronald M. Harper

Obstructive sleep apnea (OSA) is accompanied by cognitive, motor, autonomic, learning, and affective abnormalities. The putamen serves several of these functions, especially motor and autonomic behaviors, but whether global and specific sub-regions of that structure are damaged is unclear. We assessed global and regional putamen volumes in 43 recently-diagnosed, treatment-naïve OSA (age, 46.4 ± 8.8 years; 31 male) and 61 control subjects (47.6 ± 8.8 years; 39 male) using high-resolution T1-weighted images collected with a 3.0-Tesla MRI scanner. Global putamen volumes were calculated, and group differences evaluated with independent samples t-tests, as well as with analysis of covariance (covariates; age, gender, and total intracranial volume). Regional differences between groups were visualized with 3D surface morphometry-based group ratio maps. OSA subjects showed significantly higher global putamen volumes, relative to controls. Regional analyses showed putamen areas with increased and decreased tissue volumes in OSA relative to control subjects, including increases in caudal, mid-dorsal, mid-ventral portions, and ventral regions, while areas with decreased volumes appeared in rostral, mid-dorsal, medial-caudal, and mid-ventral sites. Global putamen volumes were significantly higher in the OSA subjects, but local sites showed both higher and lower volumes. The appearance of localized volume alterations points to differential hypoxic or perfusion action on glia and other tissues within the structure, and may reflect a stage in progression of injury in these newly-diagnosed patients toward the overall volume loss found in patients with chronic OSA. The regional changes may underlie some of the specific deficits in motor, autonomic, and neuropsychologic functions in OSA.

YNICL Journal 2014 Journal Article

Regional brain gray and white matter changes in perinatally HIV-infected adolescents

• Manoj K. Sarma
• Rajakumar Nagarajan
• Margaret A. Keller
• Rajesh Kumar
• Karin Nielsen-Saines
• David E. Michalik
• Jaime Deville
• Joseph A. Church

Despite the success of antiretroviral therapy (ART), perinatally infected HIV remains a major health problem worldwide. Although advance neuroimaging studies have investigated structural brain changes in HIV-infected adults, regional gray matter (GM) and white matter (WM) volume changes have not been reported in perinatally HIV-infected adolescents and young adults. In this cross-sectional study, we investigated regional GM and WM changes in 16 HIV-infected youths receiving ART (age 17.0 ± 2.9 years) compared with age-matched 14 healthy controls (age 16.3 ± 2.3 years) using magnetic resonance imaging (MRI)-based high-resolution T1-weighted images with voxel based morphometry (VBM) analyses. White matter atrophy appeared in perinatally HIV-infected youths in brain areas including the bilateral posterior corpus callosum (CC), bilateral external capsule, bilateral ventral temporal WM, mid cerebral peduncles, and basal pons over controls. Gray matter volume increase was observed in HIV-infected youths for several regions including the left superior frontal gyrus, inferior occipital gyrus, gyrus rectus, right mid cingulum, parahippocampal gyrus, bilateral inferior temporal gyrus, and middle temporal gyrus compared with controls. Global WM and GM volumes did not differ significantly between groups. These results indicate WM injury in perinatally HIV-infected youths, but the interpretation of the GM results, which appeared as increased regional volumes, is not clear. Further longitudinal studies are needed to clarify if our results represent active ongoing brain infection or toxicity from HIV treatment resulting in neuronal cell swelling and regional increased GM volume. Our findings suggest that assessment of regional GM and WM volume changes, based on VBM procedures, may be an additional measure to assess brain integrity in HIV-infected youths and to evaluate success of current ART therapy for efficacy in the brain.

EAAI Journal 2010 Journal Article

Fingerprint matching using multi-dimensional ANN

• Rajesh Kumar
• B.R. Deva Vikram

YNIMG Journal 2004 Journal Article

A method for removal of global effects from fMRI time series

• Paul M Macey
• Katherine E Macey
• Rajesh Kumar
• Ronald M Harper