Social media Annotation is the categorization and labeling of user-generated data mainly for AI applications, including media, social connections, multimodal content and descriptions, location, and comments in various forms metadata are considered valuable resources for improving the results of tasks such as semantic indexing and retrieval. But the annotation is not as easy a task as all the data available may be of low quality, difficult to interpret, personalized, or have very few images. We think that tag refinement is the most general problem and the others as specializations. The vast majority of tag refinement methods is done with images, while very few are done with videos.
1. Image Data:- Image tag refinement methods can be distinguished into two major categories, respectively following statistical modeling or data-driven approaches. The former which is statistical modeling techniques have many efficient results, but has some drawback that learning must be updated regularly as new images or terms are added, which might not seem practical in large-scale continuously evolving collections. However, the Data-driven method has a much easier application. They have been successfully used for tag ranking in social image retrieval, tag suggestion for automatic image annotation 2. Video data:- A lot of internet video work is done generally on near-duplicate detection, concept detection, or topic detectors. They observed that computation of tag relevance on the tags o the whole video would not consider the fact that some tags sonly refer to specific shots, and would result in re-ranking the same list of tags for all the shots. Thereafter they introduced various variations of the tag relevance algorithm keyframes and images to add new tags that were not originally present in the video. This model was further taken in to compute weighted tag relevance based on visual similarity, and performing an initial video tag expansion.
If you are looking forward to passing this exam or just want to get help with any other Lionbridge/Appen exam, feel free to contact us and we will make sure that all your questions are answered.
Contact Us at