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: CoRR (2022)
Keyphrases
- caption text
- image data
- false positives
- visual features
- image content
- image classification
- bounding box
- image collections
- input image
- multiscale
- single image
- image representation
- test images
- image analysis
- false negative
- image features
- pixel values
- edge detection
- low level
- image retrieval
- image pixels
- data collection
- computational complexity
- feature extraction
- image segmentation
- computer vision
- high resolution
- region of interest
- object recognition
- object categories
- similarity measure
- static images
- visual attributes
- false negative rate
- image processing