Open-Vocabulary 3D Detection via Image-level Class and Debiased Cross-modal Contrastive Learning.
Yuheng LuChenfeng XuXiaobao WeiXiaodong XieMasayoshi TomizukaKurt KeutzerShanghang ZhangPublished in: CoRR (2022)
Keyphrases
- cross modal
- image classification
- image data
- image retrieval
- image segmentation
- object detection
- perceptual information
- image content
- low level
- multiscale
- image features
- image representation
- image collections
- positive examples
- visual recognition
- visual similarity
- learning algorithm
- multi modal
- visual features
- bag of words
- similarity measure