Bidirectional Multimodal Recurrent Neural Networks with Refined Visual Features for Image Captioning.
Yanwu ShuLiyan ZhangZechao LiJinhui TangPublished in: ICIMCS (2017)
Keyphrases
- visual features
- recurrent neural networks
- image classification
- image retrieval
- image collections
- visual appearance
- image categorization
- web images
- low level
- labeled images
- visual content
- global features
- visual descriptors
- visual information
- image search
- visually similar
- semantic gap
- visual data
- low level visual features
- sample images
- image content
- bag of features
- semantic concepts
- image annotation
- neural network
- visual similarity
- low level features
- multiscale
- feed forward
- visual attributes
- input image
- key frames
- image data
- image representation
- saliency map
- cbir systems
- multi modal
- image features
- artificial neural networks
- visual patterns
- genetic algorithm
- image regions
- audio visual
- information retrieval
- automatic image annotation
- relevance feedback
- image matching
- image database