Unified Lightweight Deep Learning Frameworks for Visual Pattern Recognition Across Human-Centric and Environmental Applications
Keywords:
Visual pattern recognition, deep learning, sign language recognition, human activity analysisAbstract
Visual pattern recognition has emerged as one of the most influential research domains within computer vision and machine learning, driven by rapid advances in deep learning architectures and sensing technologies. Across diverse application areas such as human activity recognition, sign language interpretation, physiotherapy monitoring, and environmental sensing for solar energy systems, researchers increasingly face a shared set of challenges related to data scarcity, computational efficiency, robustness, and real-world deployment constraints. This article presents an integrated and theory-driven research study that synthesizes and critically analyzes state-of-the-art approaches to lightweight deep learning–based visual recognition, drawing strictly on established academic literature. By examining human-centric domains such as sign language recognition, gesture and pose analysis, and weakly supervised activity recognition alongside environmental applications including dust detection and solar panel monitoring, this work develops a unified conceptual framework for efficient visual learning. Methodological principles spanning convolutional neural networks, transformer-based models, pose estimation pipelines, and context-aware feature learning are elaborated in depth. The analysis demonstrates that despite domain differences, common architectural strategies—such as efficient model scaling, feature disentanglement, and representation optimization—play a decisive role in achieving accuracy and deployability. The results are discussed in a descriptive and comparative manner, highlighting how model design choices influence performance, generalization, and system scalability. The discussion further explores limitations, ethical considerations, and future research directions, emphasizing cross-domain transferability and the growing importance of lightweight intelligence at the edge. The study concludes that unifying human-centric and environmental vision research under shared theoretical and methodological principles can accelerate innovation and foster more resilient, inclusive, and sustainable intelligent systems.
References
Ajmal, M., Ahmad, F., Naseer, M., & Jamjoom, M. (2019). Recognizing human activities from video using weakly supervised contextual features. IEEE Access, 7, 98420–98435.
Al Abdullah, B., Amoudi, G., & Alghamdi, H. (2024). Advancements in sign language recognition: A comprehensive review and future prospects. IEEE Access.
Al-Hammadi, M., Muhammad, G., Abdul, W., Alsulaiman, M., Bencherif, M. A., Alrayes, T. S., et al. (2020). Deep learning-based approach for sign language gesture recognition with efficient hand gesture representation. IEEE Access, 8, 192527–192542.
Al Khuzayem, L., Shafi, S., Aljahdali, S., Alkhamesie, R., & Alzamzami, O. (2024). Efhamni: A deep learning-based Saudi sign language recognition application. Sensors, 24(10), 3112.
Alyami, S., Luqman, H., & Hammoudeh, M. (2024). Isolated Arabic sign language recognition using a transformer-based model and landmark keypoints. ACM Transactions on Asian and Low-Resource Language Information Processing, 23(1), 1–19.
Arrowsmith, C., Burns, D., Mak, T., Hardisty, M., & Whyne, C. (2022). Physiotherapy exercise classification with single-camera pose detection and machine learning. Sensors, 23(1), 363.
Bantupalli, K., & Xie, Y. (2018). American sign language recognition using deep learning and computer vision. In 2018 IEEE International Conference on Big Data (pp. 4896–4899). IEEE.
Banzi, J. F., & Leonard, S. (2024). Hand LightWeightNet: An optimized hand pose estimation for interactive mobile interfaces. International Journal of Electrical & Computer Engineering, 14(2).
Baytaş, İ. M., & Erdoğan, İ. (2024). Signer-independent sign language recognition with feature disentanglement. Turkish Journal of Electrical Engineering and Computer Sciences, 32(3), 420–435.
Mohammed, H. A., Baha’a, A. M., & Al-Mejibli, I. S. (2018). Smart system for dust detecting and removing from solar cells. Journal of Physics: Conference Series, 1032(1), 012055.
Nayak, S. Leveraging predictive maintenance with machine learning and IoT for operational efficiency across industries.
Onim, M. S. H., Sakif, Z. M. M., Ahnaf, A., Kabir, A., Azad, A. K., Oo, A. M. T., et al. (2022). SolNet: A convolutional neural network for detecting dust on solar panels. Energies, 16(1), 155.
Phoolwani, U. K., Sharma, T., Singh, A., & Gawre, S. K. (2020). IoT based solar panel analysis using thermal imaging. In 2020 IEEE International Students’ Conference on Electrical, Electronics and Computer Science (pp. 1–5). IEEE.
Proietti, A., Panella, M., Leccese, F., & Svezia, E. (2015). Dust detection and analysis in museum environment based on pattern recognition. Measurement, 66, 62–72.
Sandler, M., Howard, A., Zhu, M., Zhmoginov, A., & Chen, L. C. (2018). MobileNetV2: Inverted residuals and linear bottlenecks. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 4510–4520).
Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition.
Sladojevic, S., Arsenovic, M., Anderla, A., Culibrk, D., & Stefanovic, D. (2016). Deep neural networks based recognition of plant diseases by leaf image classification. Computational Intelligence and Neuroscience.
Szegedy, C., Vanhoucke, V., Ioffe, S., Shlens, J., & Wojna, Z. (2016). Rethinking the inception architecture for computer vision. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (pp. 2818–2826).
Tan, M., & Le, Q. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. In International Conference on Machine Learning (pp. 6105–6114).
Thomas, S. K., Joseph, S., Sarrop, T. S., Haris, S. B., & Roopak, R. (2018). Solar panel automated cleaning system. In 2018 International Conference on Emerging Trends and Innovations in Engineering and Technological Research (pp. 1–3). IEEE.
Zainuddin, N. F., Mohammed, M. N., Al-Zubaidi, S., & Khogali, S. I. (2019). Design and development of smart self-cleaning solar panel system. In 2019 IEEE International Conference on Automatic Control and Intelligent Systems (pp. 40–43). IEEE.
