Data Fusion and Confidence Measure in Image Feature Detection

The data fusion approaches for uncorrelated and correlated data are carried out in image feature detection, which integrate information of the same feature based on multiple methods using a statistical approach. Validity of the scheme and properties of the fused data are discussed. A confidence measure is defined, and applied to evaluate credibility of the result. The technique can be used to reduce adverse effects of individual feature detection errors, and improve the rate of pattern recognition. Examples of facial canthus detection and the corresponding confidence levels are presented.

[1]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[2]  James Llinas,et al.  An introduction to multisensor data fusion , 1997, Proc. IEEE.

[3]  S. M. Steve.  SUSAN - a new approach to low level image processing , 1997 .

[4]  Jian Yang,et al.  Generalized K-L transform based combined feature extraction , 2002, Pattern Recognit..

[5]  Anil K. Jain,et al.  Feature Selection: Evaluation, Application, and Small Sample Performance , 1997, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  Qiang Ji,et al.  Active and dynamic information fusion for facial expression understanding from image sequences , 2005, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7]  Chris H. Q. Ding,et al.  Evolving Feature Selection , 2005, IEEE Intell. Syst..

[8]  Hanseok Ko,et al.  Bayesian fusion of confidence measures for speech recognition , 2005, IEEE Signal Process. Lett..

[9]  Wen Gao,et al.  Information fusion in face identification , 2004, Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004..