Evaluating Image Blurring for Photographic Portraiture
Image blurring is a major source of image degradations that leads to loss of details. Detecting blurring region of a photo and evaluating the degree of blurring are crucial for image quality assessment. It’s more complicated for portraiture when the blurring sometimes is deliberate by photographer for visual effects. Aiming at evaluating the blurring metric of snapshot or amateur photos with faces, the paper proposed a simple and effective no-reference method of evaluating image blurring. The key idea is taking into account the artistic purpose on portraiture. The proposed method is based on the concept of Cumulative Probability of Blur Detection and pooling strategy. After computing the burring metric of pixels, the final value of global blurring evaluation is obtained with the pooling strategy according to the characteristic of human skin. Experimental results on public databases and photos collected from Internet show the proposed method can significantly improve the accuracy of objective blurring evaluation metric that have stronger correlation with subjective human assessment.
KeywordsImage quality assessment Image blurring Skin detection
- 1.Rodden, K., Wood, K.R.: Searching and organizing: how do people manage their digital photographs? In: Proceedings of the 2003 Conference on Human Factors in Computing Systems, Ft. Lauderdale, Florida, USA (2003)Google Scholar
- 4.Marziliano, P., Dufaux, F., Winkler, S., Ebrahimi, T.: A no-reference perceptual blur metric. In: Proceedings of IEEE International Conference on Image Processing, pp. 57–60 (2002)Google Scholar
- 6.Levin, A.: Blind motion deblurring using image statistics. In: Advances in Neural Information Processing Systems, vol. 19, pp. 841–848 (2007)Google Scholar
- 8.Cho, T., Paris, S., Horn, B., Freeman, W.: Blur kernel estimation using the radon transform. In: Proceedings of CVPR, pp. 241–248 (2011)Google Scholar
- 11.Narvekar, N.D., Karam, L.J.: A no-reference perceptual quality metric based on cumulative probability of blur detection. In: First International Workshop on Quality of Multimedia Experience, pp. 87–91 (2009)Google Scholar