Apr 20th, 1:00 PM - 4:00 PM
Statistical Scanner Model Validation
The process of scanning inherently introduces degradations to an image. These degradations affect the overall quality of the scanned image. Models have been derived to characterize the image degradations created by the optical system in a scanner. These models can be used to create the synthetic images. If the models accurately reflect the process of the scanner, the output of the model should match real scans. Through statistical analysis of these images, the models can be validated for accuracy. Experiments are conducted to compare the similarity of the real and synthetic images. These images are compared using several difference metrics: Hamming distance, moment features, and RST-invariant moment features, and several image description features.