ST-LDM: A Universal Framework for Text-Grounded Object Generation in Real Images.
Xiangtian XueJiasong WuYouyong KongLotfi SenhadjiHuazhong ShuPublished in: CoRR (2024)
Keyphrases
- real objects
- image data
- rigid body
- multiple objects
- image regions
- image analysis
- image database
- three dimensional
- d objects
- visual appearance
- lighting conditions
- individual objects
- image retrieval
- object recognition
- input image
- image classification
- deformable objects
- partial occlusion
- viewing angle
- objects represented
- background clutter
- information retrieval
- complex scenes
- target object
- web images
- image collections
- keypoints
- visual features
- edge detection
- object detection
- image features
- multiple modalities
- acquired images
- text generation
- computer vision