Low cost blur image detection and estimation for mobile devices

In this paper, we propose low-cost blur detection and estimation method which can be applied to improving image display on mobile phones. We detect the blurred image by constructing the Bayes discriminant function with the statistics of the gradients of the input image. Once the blurred image is detected, we extract the main ripple components from the log spectrum of the blurred image and measure the angle and width of the main ripple components to estimate the angle and extent of the blur kernel. The experimental results are provided to show the performance of the proposed blur kernel estimation and image restoration algorithm.