Improving the Performance of Convolutional Neural Networks by Fusing Low-Level Features With Different Scales in the Preceding Stage.
Xiaohong YuWei LongYanyan LiXiaoqiu ShiLin GaoPublished in: IEEE Access (2021)
Keyphrases
- low level features
- convolutional neural networks
- low level
- high level
- semantic gap
- visual features
- image retrieval
- high level semantics
- relevance feedback
- image representation
- higher level
- high level features
- semantic information
- mid level
- key frames
- convolutional network
- cbir systems
- semantic concepts
- object recognition
- high level knowledge
- automatic image annotation
- image annotation
- image database
- information retrieval
- low level visual features
- low level descriptors
- database
- visual information
- semantic content
- multimedia