MAFormer: A transformer network with multi-scale attention fusion for visual recognition.
Huixin SunYunhao WangXiaodi WangBin ZhangYing XinBaochang ZhangXianbin CaoErrui DingShumin HanPublished in: Neurocomputing (2024)
Keyphrases
- visual recognition
- multiscale
- image classification
- object recognition
- distribution network
- visual recognition tasks
- scale space
- fault diagnosis
- latent topic models
- machine learning
- category level
- image representation
- fuzzy logic
- computer vision
- coarse to fine
- local binary pattern
- image retrieval
- visual classification
- view independent
- visual categorization
- three dimensional
- image processing