Paper titled "DoppelSearch: A Novel Approach to Content-Based Video Retrieval for AI Challenge HCMC 2023"
Link: https://doi.org/10.1145/3628797.3628950
The dedicated team of students behind this innovative research includes:
Phan Nguyễn Hữu Phong – 22521090 – KHTN2022 - Main Author
Trần Đình Khoa – 22520689 – ATTN2022 – Co-author
Trần Kim Ngọc Ngân – 22520002 – KHTN2022 – Co-author
Trần Đức Lương – 19521815 – ANTN2019 – Co-author
Lê Hoàng Phúc – 20521762 – CNCL2020 – Co-author
Esteemed instructors:
Associate Professor Nguyễn Hữu Quyền
Dr. Phạm Văn Hậu
Summary:
Searching for videos, a task considered crucial in the fields of computer vision and image recognition, has found widespread applications in various sectors such as education, entertainment, security, and healthcare. However, this task encounters challenges due to the complexity of video data, instability in feature extraction methods, and differences in meaning between videos and text. In this paper, we introduce a novel content-based video search method called DoppelSearch. We leverage the CLIP (Contrastive Language-Image Pre-training) model architecture for classifying and labeling video segments, providing specific content-based video search capabilities for users. Our approach utilizes the ViT-b/32 model for feature extraction and employs feature embeddings, combined with the Faiss library, to enhance search performance. Experimental results demonstrate the high accuracy and fast retrieval times of our model, opening new possibilities for content-based video search for researchers, developers, and end-users. This paper not only introduces the applications of CLIP and ViT-b/32 models but also presents the specific feature extraction process and the use of Faiss for optimizing video search. The DoppelSearch method represents a significant advancement in the field of video search and promises diverse applications across various industries...
We express our sincere gratitude to Dr. Phạm Văn Hậu, the Head of Information Security Division, for fostering such an academically stimulating environment for our students. Additionally, we thank Dr. Nguyễn Hữu Quyền, the head of TeamQ from Wanna.W1n Club, for his continuous support, guidance, and passion for research, which has inspired our students to excel in their research endeavors and personal development. We appreciate the instructors' dedicated teaching, insightful feedback, and valuable suggestions, which have significantly contributed to the success of this research project.
The SOICT (Symposium on Information and Communication Technology) is a prestigious international scientific conference covering vital research areas such as AI Foundations and Big Data, Network Communication and Security, Image and Natural Language Processing, Software Engineering and Digital Technology, Blockchain, and Operations Research trends. The 12th SOICT conference will be held in Ho Chi Minh City, organized by the University of Information Technology and Communications - Hanoi University of Science and Technology, VNUHCM- University of Science, Laboratory Informatics, Modelling and Optimisation System (LIMOS), The French National Centre for Scientific Research (CNRS), and the Institute of Research and High Education in Mathematics (IRHÉMA). The conference will feature workshops and keynote speeches by world-class speakers. The proceedings of SOICT 2023 will be published and archived in the ACM Digital Library as part of the ACM International Conference Proceedings Series. SOICT 2023 will be indexed by DBLP, Ei Compendex, Scopus, and Clarivate Analytics Web of Science (ISI Web of Science).
Once again, congratulations to the UIT student team for their outstanding achievement in the field of scientific research. Their dedication and innovation have made significant contributions to the academic community and the advancement of knowledge in information and communication technology.
Detailed Information: https://www.facebook.com/UIT.Fanpage/posts/pfbid02w9adWr86UrYxME64oxJNBMskYvL7tFkRVWK9KCYESsZh2UJnL6YjPoUcWqX5WmXNl
Hạ Băng - Collaborator, Information Technology University Communication Team
English version: Phan Huy Hoang