Conference paper
Attention-Based image captioning using DenseNet features
Neural Information Processing, Vol.1143
26th International Conference, ICONIP 2019 (Sydney, NSW, 12/12/2019–15/12/2019)
2019
Abstract
We present an attention-based image captioning method using DenseNet features. Conventional image captioning methods depend on visual information of the whole scene to generate image captions. Such a mechanism often fails to get the information of salient objects and cannot generate semantically correct captions. We consider an attention mechanism that can focus on relevant parts of the image to generate fine-grained description of that image. We use image features from DenseNet. We conduct our experiments on the MSCOCO dataset. Our proposed method achieved 53.6, 39.8, and 29.5 on BLEU-2, 3, and 4 metrics, respectively, which are superior to the state-of-the-art methods.
Details
- Title
- Attention-Based image captioning using DenseNet features
- Authors/Creators
- M.Z. Hossain (Author/Creator) - Murdoch UniversityF. Sohel (Author/Creator) - Murdoch UniversityM.F. Shiratuddin (Author/Creator) - Murdoch UniversityH. Laga (Author/Creator) - Murdoch UniversityM. Bennamoun (Author/Creator) - The University of Western Australia
- Publication Details
- Neural Information Processing, Vol.1143
- Conference
- 26th International Conference, ICONIP 2019 (Sydney, NSW, 12/12/2019–15/12/2019)
- Identifiers
- 991005544524107891
- Murdoch Affiliation
- Information Technology, Mathematics and Statistics
- Language
- English
- Resource Type
- Conference paper
Metrics
112 Record Views