VQS: Linking Segmentations to Questions and Answers for Supervised Attention in VQA and Question-Focused Semantic Segmentation.
Chuang GanYandong LiHaoxiang LiChen SunBoqing GongPublished in: ICCV (2017)
Keyphrases
- semantic segmentation
- answer questions
- label transfer
- supervised methods
- correct answers
- weakly supervised
- question answer pairs
- street scenes
- superpixels
- community question answering
- object segmentation
- conditional random fields
- image database
- question answering
- image segmentation
- scene classification
- semi supervised
- object class
- medical images
- pascal voc
- natural images
- multiscale
- object recognition
- object categories
- probabilistic model
- supervised learning
- relation extraction
- object detection
- graph cuts
- ground truth
- long range
- saliency map
- named entities
- training data
- computer vision