MSU-Net: the multi-scale supervised U-Net for image splicing forgery localization.
Hao YuLichao SuChenwei DaiJinli WangPublished in: Pattern Anal. Appl. (2024)
Keyphrases
- object recognition
- image representation
- multiscale
- image features
- template matching
- keypoints
- natural images
- image matching
- image content
- input image
- unsupervised learning
- image classification
- test images
- image data
- edge detection
- forgery detection
- single image
- spatial color
- scale space
- digital images
- image segmentation
- segmentation method
- image regions
- high resolution
- pixel values
- multiple scales
- object localization
- learning algorithm
- low level
- image analysis
- image retrieval