Dimensional Blurring of Large Face Image Based on Structural Similarity

Main Article Content

Thitiporn Lertrusdachakul
Kasem Thiptarajan
Kanakarn Ruxpaitoon
Kulwadee Somboonviwat


     Face masking is widely used in print media, digital media and various online images for privacy and security protection. This research proposes an innovative face-blurring method focusing on large face blurring in large image size which usually requires a high degree of blur level to anonymize face. This great blur causes the face to be flat or very smooth and lack of visual dimension. Therefore, this research exploits the local structural similarity between large blurred and original images with Gaussian filter and contrast adjustment to adaptively create an opacity map for smoothly and appropriately blurring of face components. The details of main components are then very blurry and difficult to recognize personal identity while other parts of the face are less blurred. This adaptively face smoothing can improve a sense of dimensional perception and help to visually anonymize the portrait more natural resulting in the higher average value of structural similarity to the original image. The proposed method can be further applied to an aesthetically image censoring.


Download data is not yet available.

Article Details

Research Article


[1] M. Gawron and U. Boryczka, “Heterogeneous fog generated with the effect of light scattering and blur,” JACS, vol. 26, no. 2, pp. 31-44, Oct. 2019.

[2] G. Mather, “The use of image blur as a depth cue,” Perception, vol. 26, no. 9, pp. 1147-1158, Sep. 1997.

[3] J. D. Pfautz, “Depth perception in computer graphics,” Computer Laboratory, University of Cambridge, Cambridge, United Kingdom, Tech. Rep. UCAM-CL-TR-546, Sep. 2002.

[4] S. A. Cholewiak, G. D. Love, and M. S. Banks, “Creating correct blur and its effect on accommodation,” Journal of Vision, vol. 18, no. 9, pp. 1-29, Sep. 2018.

[5] B. Ham, M. Cho, and J. Ponce, “Robust guided image filtering using nonconvex potentials,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 40, no. 1, pp. 291-207, Jan. 2018, doi: 10.1109/TPAMI.2017.2669034.

[6] L. Bao, Y. Song, Q. Yang, H. Yuan, and G. Wang, “Tree filtering: Efficient structure-preserving smoothing with a minimum spanning tree,” IEEE Transactions on Image Processing, vol. 23, no. 2, pp. 555-569, Feb. 2014, doi: 10.1109/TIP.2013.2291328.

[7] Q. Liu, B. Xiong, and M. Zhang, “Adaptive sparse norm and nonlocal total variation methods for image smoothing,” Mathematical Problems in Engineering, vol. 2014, Dec. 2014, Art. no. 426125, doi: 10.1155/2014/426125.

[8] F. Zhu, Z. Liang, X. Jia, L. Zhang, and Y. Yu, “A benchmark for edge-preserving image smoothing,” IEEE Transactions on Image Processing, vol. 28, no. 7, pp. 3556-3570, Jul. 2019, doi: 10.1109/TIP.2019.2908778.

[9] Z. A. Liu, Y. K. Hou, X. T. Zhen, J. Xu, L. Shao, and M. M. Cheng. (2020). Pixel-level non-local image smoothing with objective evaluation [Online]. Available: https://csjunxu.github.io/paper/PNLS.pdf

[10] L. Xu, C. Lu, Y. Xu, and J. Jia, “Image smoothing via L0 gradient minimization,” ACM Transactions on Graphics, vol. 30, no. 6, Dec. 2011, Art. no. 174, doi: 10.1145/2024156.2024208.

[11] X. Pang, S. Zhang, J. Gu, L. Li, B. Liu, and H. Wang, “Improved L0 gradient minimization with L1 fidelity for image smoothing,” Plos One, vol. 10, no. 9, Sep. 2015, doi: 10.1371/journal.pone.0138682.

[12] S. Velusamy, R. Parihar, R. Kini, and A. Rege, “FabSoften: Face beautification via dynamic skin smoothing, guided feathering, and texture restoration,” in Proc. 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), Seattle, WA, USA, 2020, pp. 2248-2256, doi: 10.1109/CVPRW50498.2020.00273.

[13] Adobe. “Adobe Photoshop.” ADOBE.com. https://www.adobe.com/products/photoshop.html (accessed Sep. 26, 2020).

[14] Rubber Duck Labs Inc. “Privacy image editor & anonymizer with face detection.” FACEPIXELIZER.com. https://www.facepixelizer.com (accessed Sep. 26, 2020).

[15] Z. Wang, A. C. Bovik, H. R. Sheikh, and E. P. Simoncelli, “Image quality assessment: From error visibility to structural similarity,” IEEE Transactions on Image Processing, vol. 13, no. 4, pp. 600-612, Apr. 2004, doi: 10.1109/TIP.2003.819861.

[16] K. Zuiderveld, “Contrast limited adaptive histogram equalization,” in Graphic Gems IV. San Diego, CA, USA: Academic Press Professional, 1994, pp. 474-485.

[17] R. Lienhart, A. Kuranov, and V. Pisarevsky “Empirical analysis of detection cascades of boosted classifiers for rapid object detection,” in Proc. of the 25th DAGM Symposium on Pattern Recognition, Magdeburg, Germany, Sep. 2003, pp. 297-304.

[18] P. Viola and M. Jones, “Robust real-time object detection,” International Journal of Computer Vision, vol. 57, pp. 137-154, May 2004.

[19] USC-SIPI image database, University of Southern California, Nov. 2020. [Online]. Available: http://sipi.usc.edu/database

[20] Labeled faces in the wild, University of Massachusetts, Amherst, Nov. 2020. [Online]. Available: http://vis-www.cs.umass.edu/lfw

[21] Shutterstock, Inc.,“Stock images,” SHUTTERSTOCK.com. https://www.shutterstock.com/images (accessed Nov. 1, 2020).