Publications

View Google Scholar Profile

Peer-Reviewed Journal Publications (SCI/SCIE/Scopus Indexed)

1. Pseudo-labeling driven refinement of benchmark object detection datasets via analysis of learning patterns . Min Je Kim, Muhammad Munsif, Altaf Hussain, Hikmat Yar, Sung Wook Baik. Neurocomputing, 2026.
   Abstract

   This work investigates pseudo-labeling driven refinement strategies for benchmark object detection datasets by analyzing model learning patterns. The study identifies annotation inconsistencies and noisy labels through iterative training dynamics, enabling systematic dataset improvement and enhanced detection performance.

2. Quality over quantity: a data-centric survey of annotation errors in object detection datasets . Adnan Hussain, Kaleem Ullah, Muhammad Afaq, Muhammad Munsif, Altaf Hussain, Sung Wook Baik. Artificial Intelligence Review, 2026.
   Abstract

   High-quality annotations are fundamental to the success of object detection models, yet real-world datasets are often riddled with annotation errors that silently degrade performance. This data-centric survey provides a comprehensive analysis of annotation errors in object detection datasets, categorizing error types such as missing labels, incorrect bounding boxes, class confusion, and duplicate annotations. We examine how these errors propagate through the training pipeline and affect model generalization. The survey reviews existing error detection and correction methodologies, benchmarking tools, and quality assessment strategies, while highlighting the trade-off between annotation quantity and quality. Our findings underscore the critical need for rigorous data curation practices and propose future directions for building more reliable, annotation-error-aware object detection systems.

3. Action understanding in low-light and pitch-dark conditions: A comprehensive survey. Muhammad Munsif, Samee Ullah Khan, Noman Khan, Altaf Hussain, Min Je Kim, Sung Wook Baik. Engineering Applications of Artificial Intelligence, 2025.
   Abstract

   This comprehensive survey reviews deep learning methods for human action recognition under low-light and pitch-dark conditions, analyzing challenges, datasets, and state-of-the-art techniques for dark environment video analytics.

4. Hierarchical attention-based framework for enhanced prediction and optimization of organic and inorganic material synthesis. Muhammad Munsif, Altaf Hussain, Zulfiqar Ahmad Khan, Min Je Kim, Sung Wook Baik. Advanced Engineering Informatics, 2025.
   Abstract

   This paper proposes a hierarchical attention-based deep learning framework for predicting and optimizing the synthesis of organic and inorganic materials, leveraging multi-scale feature extraction for materials informatics.

5. Proximal policy optimization for collision avoidance and motion planning in autonomous vehicles. Muhammad Hijji†, Muhammad Munsif†, et al. Fractals, 2025.
   Abstract

   This work presents a mathematical modeling perspective on using proximal policy optimization (PPO) for autonomous vehicle motion planning and collision avoidance in complex environments.

6. Darkness-Adaptive Action Recognition: Leveraging Efficient Tubelet Slow-Fast Network for Industrial Applications. Muhammad Munsif, Noman Khan, Altaf Hussain, Min Je Kim, Sung Wook Baik. IEEE Transactions on Industrial Informatics, 2024.
   Abstract

   We propose an efficient Tubelet Slow-Fast Network adapted for darkness, enabling robust industrial action recognition in dark and low-light surveillance environments.

7. Contextual visual and motion salient fusion framework for action recognition in dark environments. Muhammad Munsif, Samee Ullah Khan, Noman Khan, Altaf Hussain, Min Je Kim, Sung Wook Baik. Knowledge-Based Systems, 2024.
   Abstract

   A novel contextual fusion framework combining visual appearance and motion saliency cues for robust human action recognition in challenging dark environments.

8. Attention-Based Deep Learning Framework for Action Recognition in a Dark Environment. Muhammad Munsif, Samee Ullah Khan, Noman Khan, Sung Wook Baik. Human-centric Computing and Information Sciences, 2024.
   Abstract

   This paper presents an attention-driven deep learning model for recognizing human activities in dark environments, addressing the limitations of conventional video understanding methods.

9. Optimized efficient attention-based network for facial expressions analysis in neurological health care. Muhammad Munsif, Muhammad Sajjad, Mohib Ullah, et al. Computers in Biology and Medicine, 2024.
   Abstract

   An optimized attention-based neural network for analyzing facial expressions to assist in the monitoring and healthcare of neurological disorder patients.

10. Industrial defective chips detection using deep convolutional neural network with inverse feature matching mechanism. Wasim Ullah, Samee Ullah Khan, Min Je Kim, Altaf Hussain, Muhammad Munsif, et al. Journal of Computational Design and Engineering, 2024.
    Abstract

    A deep convolutional neural network with an inverse feature matching mechanism for automated detection of defective chips in industrial manufacturing settings.

11. Optimized deep learning-based cricket activity focused network and medium scale benchmark. Waqas Ahmad, Muhammad Munsif, et al. Alexandria Engineering Journal, 2023.
    Abstract

    A deep learning pipeline optimized for fine-grained cricket action understanding, accompanied by a new medium-scale benchmark dataset for sports activity recognition.

12. CT-NET: A Novel Convolutional Transformer-Based Network for Short-Term Solar Energy Forecasting Using Climatic Information. Muhammad Munsif, Fath U Min Ullah, et al. Computer Systems Science and Engineering, 2023.
    Abstract

    A convolutional transformer network for short-term solar energy forecasting using climatic and meteorological data inputs.

13. Serious Games in Science Education. A Systematic Literature Review. Mohib Ullah, Sareer Ul Amin, Muhammad Munsif, et al. Virtual Reality and Intelligent Hardware, 2022.
    Abstract

    A systematic literature review of serious games applied in science education, evaluating their effectiveness and design patterns.

14. Automated Wheat Diseases Classification Framework Using Advanced Machine Learning Technique. Habib Khan, Ijaz Ul Haq, Muhammad Munsif, et al. Agriculture-Basel, 2022.
    Abstract

    An advanced machine learning framework for automated classification of wheat diseases from field images using deep convolutional features.

15. Deepdive: A Learning-Based Approach for Virtual Camera in Immersive Contents. Muhammad Irfan, Muhammad Munsif. Virtual Reality and Intelligent Hardware, 2022.
    Abstract

    A learning-based approach for intelligent virtual camera control in immersive VR/AR content creation.

16. An adaptive game-based learning strategy for children road safety education and practice in virtual space. Noman Khan, Khan Muhammad, Tanveer Hussain, Mansoor Nasir, Muhammad Munsif, et al. Sensors, 2021.
    Abstract

    An adaptive VR game-based learning system designed to educate children about road safety through interactive virtual environments.

Ongoing Peer-Review Submissions (SCI/SCIE Indexed)

1. View-invariant deep learning framework for action recognition in darkness. IEEE Transactions on Multimedia (Major Revision).
2. Multi-camera connected vision system with multi-view analytics: a comprehensive survey. IEEE Transactions on Big Data (Under Review).
3. Enhancing 24/7 environmental monitoring: a multimodal and illumination-resilient framework. IEEE Transactions on Image Processing (Under Review).
4. Renewable energy generation: a comprehensive survey, current challenges, and future research. Engineering Applications of AI (Under Review).
5. Multi-model structure-agnostic framework for enhanced materials discovery. Advanced Engineering Informatics (Major Revision).
6. VMAE-Time: a temporal masked autoencoding framework for self-supervised video anomaly detection. IEEE Transactions on Multimedia (Under Review).
7. Density-aware region proposal framework for enhanced object detection. Information Sciences (Under Review).
8. DATGNN: differential attention transformer-enhanced graph neural network. Engineering Applications of AI (Under Review).
9. DREAM: dual-path representation encoding and attention modeling for video retrieval. IEEE Transactions on Image Processing (Under Review).

Conference Papers (CVPR, EUVIP, ICNGC, etc.)

1. Medium Scale Benchmark for Cricket Excited Actions Understanding. Altaf Hussain, Noman Khan, Muhammad Munsif, Min Je Kim, Sung Wook Baik. IEEE CVPRW, 2024.
2. Machine Learning-Based Stability Prediction for Material Synthesis for Silicon Anodes. Altaf Hussain, Muhammad Munsif, et al. ICNGC, 2024.
3. A Survey of AI-Empowered Methods for Detecting Electricity Theft in Smart Grids. Waseem Ullah, Altaf Hussain, Muhammad Munsif, et al. ICNGC, 2023.
4. Immersive Learning: A Virtual Reality Approach to Combat Light Pollution. Younghoon Kim, Muhammad Munsif, Altaf Hussain. ICNGC, 2023.
5. Pv-anet: Attention-based network for short-term photovoltaic power forecasting. Muhammad Munsif, Habib Khan, et al. ICNGC, 2022.
6. A Lightweight Convolution Neural Network for Automatic Disasters Recognition. Muhammad Munsif, Hina Afridi, et al. IEEE EUVIP, 2022.
7. Efficient Battery's State of Charge Estimation in Energy Storage Systems. Muhammad Munsif, Noman Khan, et al. ICNGC, 2021.

Book Chapters

1. Monitoring Neurological Disorder Patients via Deep Learning Based Facial Expressions Analysis. Muhammad Munsif, Mohib Ullah, Bilal Ahmad, Muhammad Sajjad, Faouzi Alaya Cheikh. IFIP Advances in Information and Communication Technology, 2022.

Domestic Conference Presentations

1. Analyzing City-Level Population Movement in China with Graph Neural Networks. Muhammad Afaq, Adnan Hussain, Hikmat Yar, Muhammad Munsif, et al. KINGPC, 2025.
2. Surveillance Abnormal Activity Recognition Using Residual Deep Bidirectional LSTM Network. Altaf Hussain, Noman Khan, Muhammad Munsif, et al. KINGPC, 2024.
3. Industrial defective chips detection using deep convolutional neural network. Waseem Ullah, Samee Ullah Khan, Muhammad Munsif, et al. KINGPC, 2023.
4. A Diverse Viewpoint and Background Benchmark for Aerial Human Action Recognition. Muhammad Munsif, et al. KINGPC, 2023.

Preprints

1. Pseudo-labeling driven refinement of benchmark object detection datasets via analysis of learning patterns. Min Je Kim, Muhammad Munsif, et al. arXiv:2506.00997, 2025.
2. AVAR-Net: a lightweight audio-visual anomaly recognition framework with a benchmark dataset. Amjid Ali, Zulfiqar Ahmad Khan, Muhammad Munsif, et al. arXiv:2510.13630, 2025.
3. Multi-camera connected vision system with multi-view analytics: a comprehensive survey. Muhammad Munsif, Waqas Ahmad, et al. arXiv:2510.09731, 2025.