CSTrans: Correlation-guided Self-Activation Transformer for Counting Everything

• Published in: Pattern recognition Vol. 153; p. 110556
• Main Authors: Gao, Bin-Bin, Huang, Zhongyi
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.09.2024
• Subjects: Counting everything; Few-shot counting; Local dependency; Vision transformer; Counting everything; Vision transformer; Local dependency; Few-shot counting

Counting everything, also named few-shot counting, requires a model to be able to count objects with any novel (unseen) category giving few exemplar boxes. However, the existing few-shot counting methods are sub-optimal due to weak feature representation, such as the correlation between the exemplar...