Young Scientist Award
Hemraj Singh, Yonsei University, South Korea
| Hemraj Singh | |
|---|---|
| Affiliation | Yonsei University |
| Country | South Korea |
| Scopus ID | 58072710400 |
| Documents | 9 |
| Citations | 78 |
| h-index | 5 |
| Subject Area | Computer Science |
| Event | Global Scholar Awards |
| ORCID | 0000-0002-7110-5147 |
Hemraj Singh is a computer science researcher whose published work centers on computer vision, video understanding, deep learning, and efficient visual recognition. His research addresses video salient object detection, video shadow detection, facial expression recognition, lightweight neural architectures, and resource-aware visual computing. His publication record includes studies of separable convolution, deformable fusion, geometric multi-scale learning, and edge-preserving network design. These themes place his research within contemporary efforts to develop accurate and computationally efficient vision systems for complex video environments and Internet of Things applications.[5]
Abstract
Hemraj Singh is a computer science researcher working in computer vision, video analysis, deep learning, and efficient visual recognition. His research includes video salient object detection, shadow detection, facial expression recognition, and lightweight neural networks for resource-constrained applications. His publications investigate separable convolution, deformable fusion, geometric multi-scale representation, contrastive learning, and edge-preserving feature extraction. These studies address challenges involving motion, occlusion, clutter, computational cost, and visual complexity in video data. His research profile demonstrates sustained engagement with practical artificial intelligence methods for intelligent visual systems, particularly where accuracy, efficiency, and deployment considerations must be balanced in real-world environments and applications. [5]
Keywords
Hemraj Singh’s publication-oriented keywords include video salient object detection, video shadow detection, facial expression recognition, computer vision, deep learning, lightweight neural networks, separable convolution, dilated convolution, deformable fusion, geometric multi-scale learning, pixel-level contrastive learning, spatiotemporal features, edge-preserving networks, Internet of Things, visual recognition, image processing, and resource-efficient artificial intelligence. [3]
Introduction
Hemraj Singh’s research addresses computer vision problems requiring reliable interpretation of dynamic visual information. His publications focus particularly on video salient object detection, where models must identify visually important regions while handling motion, appearance variation, occlusion, blur, and complex backgrounds. His work also considers computational efficiency, reflecting the needs of practical intelligent systems and constrained deployment environments. [2]
Research Profile
Hemraj Singh’s research profile is situated within artificial intelligence and computer vision, with emphasis on video processing and deep neural network design. His documented work includes lightweight architectures for salient object detection, video shadow analysis, and facial expression recognition. The research combines spatial, temporal, geometric, and semantic representations to improve visual understanding while maintaining computational practicality. [1]
Research Contributions
Hemraj Singh’s contributions include efficient separable convolution networks for video salient object detection, deformable separable fusion for visual feature integration, and geometric multi-scale pixel-level contrastive learning for video analysis. His work on HSNet further addresses video shadow detection through hierarchical separable processing and edge-preserving representations. Together, these studies emphasize efficient feature extraction, robust spatial-temporal modeling, and practical computer vision. [2] [3] [4]
Publications
Hemraj Singh has published research including Novel Dilated Separable Convolution Networks for Efficient Video Salient Object Detection in the Wild, DSFNet: Video Salient Object Detection Using a Novel Lightweight Deformable Separable Fusion Network, DMFNet: geometric multi-scale pixel-level contrastive learning for video salient object detection, HSNet: A Novel Edge-Preserving Hierarchical Separable Network for Video Shadow Detection, DSNet: Efficient Lightweight Model for Video Salient Object Detection for IoT and WoT Applications, and DMSNet: A Lightweight and Efficient Facial Expression Recognition Model for IoT and WoT Applications. [1] [3] [4]
Research Impact
Hemraj Singh’s research contributes to computer vision by addressing the practical balance between recognition quality and computational efficiency. His studies investigate lightweight architectures applicable to video processing and Internet of Things environments, while other work targets challenging visual conditions such as shadows, clutter, motion, and scale variation. The combination of these themes supports broader development of deployable artificial intelligence systems. [2] [4]
Award Suitability
Hemraj Singh’s publication activity and research themes provide a relevant basis for consideration for a Young Scientist Award in computer science. His work demonstrates engagement with contemporary artificial intelligence challenges, particularly efficient deep learning, video understanding, salient object detection, shadow detection, and visual recognition. The documented publication record also reflects continued research development across related computer vision problems. [1] [5]
Conclusion
Hemraj Singh represents an emerging research profile in computer science focused on computer vision, video processing, and efficient deep learning. His publications address salient object detection, shadow detection, facial expression recognition, and lightweight model development. The consistent emphasis on efficient visual representations and practical deployment establishes a coherent research direction within contemporary artificial intelligence and intelligent video analysis. [1] [4]
External Links
References
- DBLP. (n.d.). Hemraj Singh: Bibliographic profile and publications. DBLP Computer Science Bibliography.
https://dblp.org/pid/338/0146.html - Singh, H., Verma, M., & Cheruku, R. (2025). DMFNet: Geometric multi-scale pixel-level contrastive learning for video salient object detection. International Journal of Multimedia Information Retrieval, 14, 12. https://doi.org/10.1007/s13735-025-00361-z
- Singh, H., Verma, M., & Cheruku, R. (2023). Novel Dilated Separable Convolution Networks for Efficient Video Salient Object Detection in the Wild. IEEE Transactions on Instrumentation and Measurement, 72. https://doi.org/10.1109/TIM.2023.3302911
- Singh, H., Verma, M., & Cheruku, R. (2025). HSNet: A Novel Edge-Preserving Hierarchical Separable Network for Video Shadow Detection. Circuits, Systems, and Signal Processing, 44, 3983–4012. https://doi.org/10.1007/s00034-024-02983-w
- Singh, H., Baithi, P., & Bashir, Z. (2025). DMSNet: A Lightweight and Efficient Facial Expression Recognition Model for IoT and WoT Applications. WWW Companion Volume, 2531–2535. https://doi.org/10.1145/3701716.3717576
