Abstract
Content-based image retrieval is quickly becoming the most common method of searching vast databases for images, giving researchers a lot of room to develop new techniques and systems. Likewise, another common application in the field of computer vision is content-based visual information retrieval. For image and video retrieval, text-based search and Web-based image reranking have been the most common methods. Though Content Based Video Systems have improved in accuracy over time, they still fall short in interactive search. The use of these approaches has exposed shortcomings such as noisy data and inaccuracy, which often result in the showing of irrelevant images or videos. The authors of the proposed study integrate image and visual data to improve the precision of the retrieved results for both photographs and videos. In response to a user's query, this study investigates alternative ways for fetching high-quality photos and related videos.
Highlights
The search results obtained from the text-based image search can be further refined using a visual re-ranking process, improving the accuracy of the text-based image search ranking
Because there are no known datasets for content-based image retrieval, Liu et al [28] use Corel datasets to evaluate the proposed model (MSD) for image retrieval (CBIR)
We conducted a thorough assessment of the literature on various CBIR and image representation techniques
Summary
Visual search has gained a lot of traction This technology, in its most basic form, provides for text-based query search within an image database. This methodology is quickly utilized for product recognition [1], [2], and location recognition [3], [4], as well as commercial applications [5], [6]. The search results obtained from the text-based image search can be further refined using a visual re-ranking process, improving the accuracy of the text-based image search ranking. Popular search engines such as Google, Yahoo, and Bing are built on the concept of retrieving photos in response to user requests utilizing annotated labels added to photos. 2] How can we integrate data based on features that can be compared directly to text, image, and video content?
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call