Attention Focused Spatial Pyramid Pooling for Boxless Action Recognition in Still Images.
Weijiang FengXiang ZhangXuhui HuangZhigang LuoPublished in: ICANN (2) (2017)
Keyphrases
- action recognition
- spatial pyramid
- bag of features
- bag of words
- image classification
- mid level
- human actions
- three dimensional
- computer vision
- object recognition
- image features
- visual words
- input image
- object classification
- test images
- image representation
- lighting conditions
- image matching
- image collections
- segmentation algorithm
- visual features
- feature points
- human activities
- image retrieval
- matching scheme
- image segmentation
- machine learning
- bounding box
- segmentation method
- image registration
- image sequences