Fusing Global and Local Features for Generalized AI-Synthesized Image Detection.
Yan JuShan JiaLipeng KeHongfei XueKoki NaganoSiwei LyuPublished in: ICIP (2022)
Keyphrases
- image features
- global features
- extracting features
- image data
- low level
- false positives
- extracted features
- input image
- cluttered scenes
- image content
- fusion method
- invariant features
- original images
- global structure
- adaboost classifier
- image classification
- multiscale
- feature representation
- image pixels
- image regions
- automatic image orientation
- single image
- feature extraction
- feature vectors
- segmentation method
- sample images
- textural features
- test images
- feature values
- bounding box
- feature set
- edge detection
- artificial intelligence
- texture and shape features
- global shape
- similarity measure
- image analysis
- feature points
- image representation
- spatial information
- detection method
- high resolution
- object detectors
- feature descriptors
- image structure
- spatial relationships
- keypoints
- image segmentation
- image matching
- detection algorithm
- geometric constraints
- hilbert scan
- computer vision
- synthesized images
- image set
- image retrieval