Annotation redundancy means the existence of duplicate or unnecessary annotation information in annotation results (e.g., repeated annotation of the same target, redundant attribute annotation), which increases data storage costs and model training interference, requiring cleanup of redundant information.
