Early Academic Pursuits

Mr. Snehashis Majhi embarked on his academic journey with a Bachelor of Technology in Computer Science and Engineering from the National Institute of Science and Technology, Berhampur, India. During this time, he delved into the realm of computer vision with his thesis titled "Design and Development of CBIR system using Classification Confidence," under the guidance of Dr. Jatindra Kumar Dash. This early exposure laid the foundation for his future endeavors in research and development.

Professional Endeavors

Mr. Majhi's professional journey commenced with an internship at INRIA, Sophia Antipolis Cedex BP 93, France, where he contributed to the development of a deep learning pipeline for human activity detection in smart home scenarios. This experience provided him with practical insights into real-world applications of deep learning technologies.

Following his internship, Majhi ventured into the field of glaucoma detection during his tenure as a Research Fellow at MNIT Jaipur, India. Here, he developed a deep convolutional neural network for detecting glaucoma diseases in fundus images, showcasing his versatility in applying deep learning techniques to healthcare domains.

Currently serving as a Research Scholar at INRIA Sophia Antipolis, France, Majhi's primary focus lies in developing human-scene centric video abnormality detection methods for smart-city surveillance systems. His research involves collaboration with Toyota Motor Europe and Woven Planet, demonstrating his ability to work on interdisciplinary projects with industry partners.

Contributions and Research Focus

Mr. Majhi's research contributions span various aspects of computer vision and deep learning. Noteworthy among his works is the development of the Human-Scene Network (HSN) and the Outlier-Embedded Cross Temporal Scale Transformer (OE-CTST) for video anomaly detection. These innovations address critical challenges in surveillance systems, showcasing Majhi's expertise in designing novel architectures and algorithms for complex tasks.

His research endeavors also encompass weakly-supervised learning techniques and optimization methods tailored for anomaly detection, underscoring his commitment to advancing the state-of-the-art in video analysis.

Accolades and Recognition

Mr. Majhi's contributions have been recognized through publications in reputable conferences and journals, including IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) and International Conference on Neural Information Processing (ICONIP). His papers have received commendable rankings, reflecting the significance and impact of his research in the scientific community.

Impact and Influence

Through his research and academic pursuits, Mr. Majhi has made significant contributions to the fields of computer vision and machine learning. His innovative approaches to video anomaly detection have the potential to enhance surveillance systems' effectiveness, thereby contributing to the advancement of public safety and security measures.

Legacy and Future Contributions

Mr. Majhi's legacy lies in his pioneering work in video anomaly detection and deep learning applications. His research serves as a stepping stone for future advancements in smart-city surveillance, healthcare diagnostics, and beyond. As he continues his academic and professional journey, Majhi is poised to make further contributions to the field, driving innovation and addressing societal challenges through cutting-edge technologies.

Citations

• Citations   97
• h-index       5
• i10-index   4

Notable Publication

• OE-CTST: Outlier-Embedded Cross Temporal Scale Transformer for Weakly-Supervised Video Anomaly Detection , 2024
• Human-Scene Network: A novel baseline with self-rectifying loss for weakly supervised video anomaly detection , 2 (2024)
• TrichANet: An Attentive Network for Trichogramma Classification , 1 (2023)
• Feature Modulating Two-Stream Deep Convolutional Neural Network for Glaucoma Detection in Fundus Images, 4, (2022)
• DAM: Dissimilarity Attention Module for Weakly-supervised Video Anomaly Detection, 10, (2021)