Face Annotation with Weakly Labeled Web Facial Images

Abstract

Auto Face annotation is playing an important role in Many real-world applications. Search-based face (SBFA) annotation is aim to begin the automated face annotation task by employ the Content Based Image Retrieval (CBIR) techniques. CBIR is used as a query by image content and automatically detect the faces from photo, assign that face to corresponding human name. Searching and mining are massive weakly labeled facial images are available in World Wide Web. Weakly labeled facial images are often noisy and incomplete. To overcome this trouble move onto the unsupervised label refinement (ULR) and Clustering based approximation techniques (CBA), to improve the accuracy, performance and scalability. By combining textual and visual features, it manages to pick “good” features that reflect users’ perception, and therefore is effective for search.