With the analysis of the features of image edge based on the defocused model of optical imaging system, a blur estimation and detection method for out-of-focus images is proposed. The essential idea is to estimate the parameter of the point spread function, which reflects the blurriness of image. Based on the notion, the proposed method estimates the parameter values by different straight edges in the image, and the parameter distribution is used to measure the image blurriness. Then it can determine whether an image is blurred or not by comparing with a predetermined threshold. Experiment results show that the proposed blur metric is highly correlated to subjective visual perception, and it can be implemented to estimate and detect the blurriness for out-of-focus images with different scenes.
Science & Technology
Computer Science, Information Systems
; Computer Science, Software Engineering
; Computer Science, Theory & Methods
; Engineering, Electrical & Electronic