Blind Measurement of Image Blur for Vision-Based Applications
This chapter presents a novel metric for image quality assessment from a single image. The key idea is to estimate the point spread function (PSF) from the line spread function (LSF), whereas the LSF is constructed from edge information. It is proven that an edge point corresponds to the local maximal gradient in a blurred image, and therefore edges can be extracted from blurred images by conventional edge detectors. To achieve high accuracy, local Radon transform is implemented and a number of LSFs are extracted from each edge. The experimental results on a variety of synthetic and real blurred images validate the proposed method. To improve the system efficiency, a criterion for edge sharpness is further proposed and only the edge points from sharp edges are selected for extracting the LSF without using all edge information. The effects of nearby edges on the selected edge feature and the resultant LSF are analyzed, and two constrains are proposed to determine appropriate LSFs. The experimental results demonstrate the accuracy and efficiency of the proposed paradigm. This scheme has fast speed and can be served in blind image quality evaluation for real-time automatic machine-vision-based applications.
KeywordsPoint Spread Function Edge Point Step Edge Image Quality Assessment Radon Transform
Unable to display preview. Download preview PDF.
- Andrews, H.C., Hunt, B.R.: Digital Image Restoration. Prentice Hall, Englewood Cliffs (1977)Google Scholar
- Beck, P.V.: Edge-based Image Representation and Coding. Ph.D. dissertation, Delft University of Technology, The Netherlands (1995)Google Scholar
- Bones, P.J., Bretschneider, T., Forne, C.J., Millane, R.P., McNeill, S.J.: Tomographic blur identification using image edges. In: Proc. SPIE, vol. 4123, pp. 133–141 (2000)Google Scholar
- Li, X.: Blind image quality assessment. In: Proc. IEEE Int. Conf. Image Processing, NY, pp. I-449–452 (2002)Google Scholar
- Luxen, M., Forstner, W.: Characterizing image quality: blind estimation of the point spread function from a single image. In: Proc. Photogrammetric Computer Vision, Austria, pp. A205–A210 (2002)Google Scholar
- Sezan, M.I., Pavlovic, G., Tekalp, A.M., Erdem, A.T.: On modeling the focus blur in image restoration. In: Proc. IEEE Int. Conf Acoust., Speech, and Signal Proc., Toronto, pp. 2485–2488 (1991)Google Scholar
- Subbarao, M., Tyan, J.-K.: The optimal focus measure for passive autofocusing and depth-from-focus. In: Proc. SPIE Int. Symp., Videometrics V, vol. 2598, pp. 89–99 (1995)Google Scholar
- Wolin, D., Johnson, K., Kipman, Y.: The importance of objective analysis in image quality evaluation. In: Proc. Int. Conf. Digital Printing Technologies, pp. 603–606 (1998)Google Scholar
- Wu, S., Chen, L., Lin, W., Jiang, L., Xiong, W., Ong, S.H.: An objective out-of-focus blur measurement. In: Proc. 5th Int. Conf. Inform., Commun. & Signal Processing, Thailand, pp. 334–338 (2005)Google Scholar