MAT: A Multimodal Attentive Translator for Image Captioning.
Chang LiuFuchun SunChanghu WangFeng WangAlan L. YuillePublished in: IJCAI (2017)
Keyphrases
- image data
- image features
- single image
- input image
- image classification
- template matching
- region of interest
- image analysis
- image matching
- low level
- image content
- image retrieval
- image representation
- multimodal image registration
- image segmentation
- edge detection
- multiscale
- image collections
- image regions
- hough transform
- image pixels
- pixel values
- lighting conditions
- image quality
- test images
- mutual information
- optical flow
- similarity measure
- face recognition