Blurry faces are sometimes recognized better than high-resolution faces

M. Kwon, A. Kalia, G. Legge
2010 Journal of Vision  
BACKGROUND RESEARCH HYPOTHESIS METHODS RESULTS CONCLUSIONS The authors thank the NIH for its financial support for this study under grant R01 EY002934. Images: 256 x 256 gray scale (8 bits) images subtending about 11º (Max Planck Institute face database) Low-pass filter: The images were blurred using a 50 th order FIR (Finite Impulse Response) low-pass spatial-frequency filter. The filter function is:
doi:10.1167/8.6.898 fatcat:2beenesnhfccfk5q6vnqpmjjka