DCT based algorithm for blurred regions determination in digital images

One of the common irregularities in digital images is blur, which is usually caused by the object motion, camera shake or out of focus. In this paper we present an algorithm for detecting images with blur, or blurred regions within an image. Blur is detected by analyzing DCT coefficients and introducing function and threshold for classification of appropriately sized sub-regions as blurred or non-blurred. Additionally, algorithm has a component for blur detection on pixel level. Our proposed algorithm exhibited superior performance compared to one of DCT based algorithms from literature.