期刊名称：International Journal of Mechatronics, Electrical and Computer Technology
印刷版ISSN：2305-0543
出版年度：2017
卷号：7
期号：24
页码：3344-3353
出版社：Austrian E-Journals of Universal Scientific Organization
摘要：Face Annotation is a note or description added to the image for better understanding. Also it can help to improve better search due to detailed description. If this annotation technique is used in video that can help in better searching of videos. The goal is to annotate unseen faces in videos with the words that best describe the image. Initially the database containing images and description mapping of that image will be gathered. Later videos that need to be processed will be considered. These videos will be converted to frames. This frame will act as images. These images will be processed with the existing database. If the faces are matched then it will be considered with the matching annotation. The matching results will produce thee matching annotation or null (the images that are not matched). Further training can be provided by the later result.The problem of naming can be traced back to name face association, where the goal is to align the observed faces with a given set of names in videos. Our proposed system give the Face candidate retrieval by name Automated video indexing by the person’s name Automated creation of face-name correspondences database from thousands of hours of news videos. Use of Annotations has increased in images by adding Videos can also use this approach for associating face-name for videos can be a approach for better video searching. It will help for users to search desired videos, eg. News videos. Also systems with manual caption exist. If such system gets implemented then captions can get added automatically. Automatic tagging of people in videos will improve the search results. It can be further enhanced by considering different parameters like image background and other parameters for providing better description.
关键词：Face Annotations; social network; Face recognition; unconstrained web videos mining; unsupervised.
