Performance assessment of face recognition using super-resolution

Recognition rate of face recognition algorithms is dependent on the resolution of the imagery, specifically the number of pixels contained within the face. Using a sequence of frames from low-resolution videos, super-resolution image reconstruction can form a higher resolution image, aiding the face recognition stage for improved performance. In this work, images from a video database of moving faces and people are used to assess the performance improvement of face recognition using super-resolution.

[1]  Hyeonjoon Moon,et al.  The FERET verification testing protocol for face recognition algorithms , 1998, Proceedings Third IEEE International Conference on Automatic Face and Gesture Recognition.

[2]  Ronald G Driggers,et al.  Superresolution image reconstruction from a sequence of aliased imagery. , 2006, Applied optics.

[3]  S. Susan Young,et al.  Super-resolution image reconstruction from a sequence of aliased imagery , 2005, SPIE Defense + Commercial Sensing.

[4]  Bruce A. Draper,et al.  The CSU Face Identification Evaluation System: Its Purpose, Features, and Structure , 2003, ICVS.

[5]  Xiaoming Liu,et al.  Multi-Frame Super-Resolution for Face Recognition , 2007, 2007 First IEEE International Conference on Biometrics: Theory, Applications, and Systems.

[6]  Luuk J. Spreeuwers,et al.  The Effect of Image Resolution on the Performance of a Face Recognition System , 2006, 2006 9th International Conference on Control, Automation, Robotics and Vision.

[7]  Alice J. O'Toole,et al.  A video database of moving faces and people , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[8]  P. Jonathon Phillips,et al.  Face Recognition Vendor Test 2002 Performance Metrics , 2003, AVBPA.