Enhanced AI-Based Image and Video Retrieval System Using CLIP and Hybrid Semantic Indexing
International Journal on Advanced Computer Engineering and Communication Technology · 22.12.2025 · 14 (3s) · 272-278
DOI: 10.65521/ijacect.v14i3s.1631 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal on Advanced Computer Engineering and Communication Technology, online erschienen am 22.12.2025. Verfasst von Anuja Bele und 4 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-ND 4.0).
Abstract (Originalsprache)
Finding pictures and movies fast and precisely has become more crucial due to the explosive growth of multimedia content. In order to improve the efficiency and semantic significance of that procedure, this research presents a sophisticated AI-powered retrieval system. The system supports both text-based and image-based searches by combining Facebook AI Similarity Search (FAISS) [10][11]with Contrastive Language–Image Pre-training (CLIP). It produces more accurate results by enabling configurable weighting and removing irrelevant or negatively associated suggestions. During video retrieval, the system extracts individual frames using FFmpeg and indexes them using FAISS for frame-level similarity matching. With Precision@5 of 92.8%, Recall@10 of 89.1%, and an average query time of just 0.45 seconds, the method achieves remarkable performance. Recent developments in multimodal video processing[25],[26] and CLIP optimization[24] are combined to increase efficiency even more. All things considered, our approach offers a scalable and useful foundation for high semantic comprehension in real-time multimedia retrieval.
Zitieren
APA
Bele, A., Lawrence, R., Butle, D., Gupta, D. K., & Hiwanj, H. (2025). Enhanced AI-Based Image and Video Retrieval System Using CLIP and Hybrid Semantic Indexing. International Journal on Advanced Computer Engineering and Communication Technology, 14(3s), 272-278. https://doi.org/10.65521/ijacect.v14i3s.1631
Vancouver
Bele A, Lawrence R, Butle D, Gupta DK, Hiwanj H. Enhanced AI-Based Image and Video Retrieval System Using CLIP and Hybrid Semantic Indexing. International Journal on Advanced Computer Engineering and Communication Technology. 2025;14(3s):272-278. doi:10.65521/ijacect.v14i3s.1631
BibTeX
@article{Bele2025,
author = {Bele, Anuja and Lawrence, Ryan and Butle, Darshan and Gupta, Dr. Kapil and Hiwanj, Himanshu},
title = {Enhanced AI-Based Image and Video Retrieval System Using CLIP and Hybrid Semantic Indexing},
journal = {International Journal on Advanced Computer Engineering and Communication Technology},
year = {2025},
volume = {14},
number = {3s},
pages = {272--278},
doi = {10.65521/ijacect.v14i3s.1631}
}
EUVIMED