ECCV Caption: Correcting False Negatives by Collecting Machine-and-Human-verified Image-Caption Associations for MS-COCO.
Sanghyuk ChunWonjae KimSong ParkMinsuk ChangSeong Joon OhPublished in: ECCV (8) (2022)
Keyphrases
- caption text
- image data
- image retrieval
- bounding box
- image content
- image features
- false positives
- false negative
- image classification
- visual features
- input image
- multiscale
- image representation
- single image
- image analysis
- image segmentation
- association rules
- image collections
- edge detection
- test images
- video retrieval
- image pixels
- high resolution
- static images
- data collection
- region based image
- visual attributes
- low false positive rate
- false negative rate