Abhijit Boruah
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Abhijit Boruah has a total of 59 co-authors. Here are the top co-authors by number of shared publications:
Nayan M. Kakoty
Tezpur University,School of Engineering,Assam,India
4 shared publications- Nayan M. Kakoty
Tezpur University
3 shared publications - Rabinder Kumar Prasad
Department of CSE, DUIET, Dibrugarh University, Dibrugarh, Assam, India
3 shared publications Tazid Ali
Dibrugarh University,Faculty of Science and Engineering,Assam,India
3 shared publications- Debajit Sarma
Department of ECE, National Institute of Technology Rourkela, Odisha, India
2 shared publications - Sudipta Majumder
Department of CSE, DUIET, Dibrugarh University, Dibrugarh, India
2 shared publications
Latest Publications
- 21 July 2026
- Published by Springer Nature
- p. 97-114
Semantic Segmentation and Scene Understanding Using RGBD Data
- 21 July 2026
- Published by Springer Nature
- p. 97-114
Semantic Segmentation and Scene Understanding Using RGBD Data
- 10 January 2026
- Published by Springer Nature
- p. 45-58
AI-Driven Secure Vehicle Verification for Smart Transportation: Multi-factor Authentication and Anomaly Detection
- 10 January 2026
- Published by Springer Nature
- p. 45-58
AI-Driven Secure Vehicle Verification for Smart Transportation: Multi-factor Authentication and Anomaly Detection
- Open Access
- 29 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 1758-1779
M k -Means: Improving Partition-Based Clustering With Optimized Centroid Initialization
The k-means algorithm is a well-known technique in partition-based clustering, drawing plenty of attention by researchers over the years. It identifies distinct, non-overlapping patterns by assigning each data point to a specific cluster using a minimum squared distance approach. However, its’ effectiveness heavily depends on the selection of initial centroid values, which are chosen randomly. Despite numerous attempts by different researchers to address this limitation, there remains significant room for improvement in the effective initialization of centroids. Here, we introduce a modified k-means, that is,Mk-means method to precisely detect all clusters in a given dataset by utilizing as objective function referred to as the sum of squared errors. This method utilizes the partial sum of the squared errors of each cluster to improve cluster identification. The efficacy of the Mk-means method is evaluated using 15 synthetic benchmark datasets and 15 real-world benchmark datasets obtained from the UCI repository. Performance comparisons were conducted against several clustering algorithms, including the traditional k-means, FCM, k-means++, and more recent approaches such as IKMN, t-k-means, fast-t-k-means, fast-t-k-means++, and Global-k-means++. The results demonstrate that Mk-means outperforms these methods in terms of external validity indices- Adjusted Rand Index (ARI), Normalized Mutual Information (NMI), and Purity, as well as sum of squared error (SSE), an internal validity index. These findings highlight that Mk-means not only achieves more accurate clustering but also reduces within-cluster variance compared to its counterparts. The source code for the Mk-means algorithm is publicly available.
- 29 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 1758-1779
Open AccessM k -Means: Improving Partition-Based Clustering With Optimized Centroid Initialization
The k-means algorithm is a well-known technique in partition-based clustering, drawing plenty of attention by researchers over the years. It identifies distinct, non-overlapping patterns by assigning each data point to a specific cluster using a minimum squared distance approach. However, its’ effectiveness heavily depends on the selection of initial centroid values, which are chosen randomly. Despite numerous attempts by different researchers to address this limitation, there remains significant room for improvement in the effective initialization of centroids. Here, we introduce a modified k-means, that is,Mk-means method to precisely detect all clusters in a given dataset by utilizing as objective function referred to as the sum of squared errors. This method utilizes the partial sum of the squared errors of each cluster to improve cluster identification. The efficacy of the Mk-means method is evaluated using 15 synthetic benchmark datasets and 15 real-world benchmark datasets obtained from the UCI repository. Performance comparisons were conducted against several clustering algorithms, including the traditional k-means, FCM, k-means++, and more recent approaches such as IKMN, t-k-means, fast-t-k-means, fast-t-k-means++, and Global-k-means++. The results demonstrate that Mk-means outperforms these methods in terms of external validity indices- Adjusted Rand Index (ARI), Normalized Mutual Information (NMI), and Purity, as well as sum of squared error (SSE), an internal validity index. These findings highlight that Mk-means not only achieves more accurate clustering but also reduces within-cluster variance compared to its counterparts. The source code for the Mk-means algorithm is publicly available.
- 10 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Semantic Spatial Object Interpretation for Indoor Environments using Sensor Fusion
Scene understanding is a critical component of vision-based decision-making systems, often requiring an effective interpretation of spatial relationships among objects in the environment. This work prioritizes spatial object interpretation by implementing an indoor scene understanding framework that integrates sensor fusion, combining 2D LiDAR and RGB cameras. A proximity-based 2D-3D object alignment approach is proposed to extract spatial semantics and visually interpret indoor environments with enhanced detail. This method incorporates spatial relationships, deducing relative object positions, distances, and hierarchical arrangements to provide a comprehensive understanding of indoor scenes. The model achieved a high precision (0.97) and recall (0.97) for object detection and segmentation, alongside a mean average precision (mAP50) of 0.98. Quantitative results demonstrated superior segmentation and spatial relationship deduction, with distances between objects calculated with an accuracy of $\mathbf{\pm 0. 0 5 ~ m}$, facilitating semantic segmentation that offers meaningful insights into object positioning within indoor environments.- 10 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-6
Semantic Spatial Object Interpretation for Indoor Environments using Sensor Fusion
Scene understanding is a critical component of vision-based decision-making systems, often requiring an effective interpretation of spatial relationships among objects in the environment. This work prioritizes spatial object interpretation by implementing an indoor scene understanding framework that integrates sensor fusion, combining 2D LiDAR and RGB cameras. A proximity-based 2D-3D object alignment approach is proposed to extract spatial semantics and visually interpret indoor environments with enhanced detail. This method incorporates spatial relationships, deducing relative object positions, distances, and hierarchical arrangements to provide a comprehensive understanding of indoor scenes. The model achieved a high precision (0.97) and recall (0.97) for object detection and segmentation, alongside a mean average precision (mAP50) of 0.98. Quantitative results demonstrated superior segmentation and spatial relationship deduction, with distances between objects calculated with an accuracy of $\mathbf{\pm 0. 0 5 ~ m}$, facilitating semantic segmentation that offers meaningful insights into object positioning within indoor environments.- 8 December 2023
- Published by Springer Nature in Soft Computing
- Vol. 28 (7-8) , 6129-6151
Enhancing visionless object recognition on grasp using ontology: the OntOGrasp framework
- 8 December 2023
- Published by Springer Nature in Soft Computing
- Vol. 28 (7-8) , 6129-6151
Enhancing visionless object recognition on grasp using ontology: the OntOGrasp framework
- 24 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 246-250
A Contactless Control Mechanism for Computerized Systems using Hand Gestures
This article presents a systematic way to develop a hand-sign recognition system aimed at detecting and interpreting user hand-signs for specific functionalities, enabling an easy contactless user experience. The process unfolds through several sequential steps. Firstly, the data collection phase involves employing a custom camera integrated with OpenCV, capturing a diverse range of hand-signs crucial for model training. Subsequently, the model training step encompasses extracting hand landmark data using mediapipe and leveraging scikit-learn in conjunction with a random forest/SVM classifier. Once the model is trained, a classifier is assigned to recognize distinct hand-signs, achieved by mapping gestures to keyboard shortcuts. Lastly, the system’s functionality is encapsulated within a tkinter GUI, yielding a cohesive user experience. This approach achieves efficient and accurate hand-sign recognition, enabling seamless interaction and activation of functions through gesture-based input in a contactless manner.
- 24 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 246-250
A Contactless Control Mechanism for Computerized Systems using Hand Gestures
This article presents a systematic way to develop a hand-sign recognition system aimed at detecting and interpreting user hand-signs for specific functionalities, enabling an easy contactless user experience. The process unfolds through several sequential steps. Firstly, the data collection phase involves employing a custom camera integrated with OpenCV, capturing a diverse range of hand-signs crucial for model training. Subsequently, the model training step encompasses extracting hand landmark data using mediapipe and leveraging scikit-learn in conjunction with a random forest/SVM classifier. Once the model is trained, a classifier is assigned to recognize distinct hand-signs, achieved by mapping gestures to keyboard shortcuts. Lastly, the system’s functionality is encapsulated within a tkinter GUI, yielding a cohesive user experience. This approach achieves efficient and accurate hand-sign recognition, enabling seamless interaction and activation of functions through gesture-based input in a contactless manner.
- Open Access
- 31 July 2023
- Published by ACCENTS in International Journal of Advanced Technology and Engineering Exploration
- Vol. 10 (104)
Investigating the performance vulnerability of AODV protocol of IoT network under SYN-flood attack
- 31 July 2023
- Published by ACCENTS in International Journal of Advanced Technology and Engineering Exploration
- Vol. 10 (104)
Open AccessInvestigating the performance vulnerability of AODV protocol of IoT network under SYN-flood attack
- 17 February 2023
- Published by Springer Nature
- p. 927-937
Real-Time Hand Gesture Recognition Using Indian Sign Language
- 17 February 2023
- Published by Springer Nature
- p. 927-937
Real-Time Hand Gesture Recognition Using Indian Sign Language
- Top Cited PapersOpen Access
- 30 January 2023
- Published by Elsevier in Procedia Computer Science
- Vol. 218, 1384-1393
Real-time Assamese Sign Language Recognition using MediaPipe and Deep Learning
- 30 January 2023
- Published by Elsevier in Procedia Computer Science
- Vol. 218, 1384-1393
Top Cited PapersOpen AccessReal-time Assamese Sign Language Recognition using MediaPipe and Deep Learning