A multiple-baseline stereo for precise human face acquisition.
PATTERN RECOGNITION LETTERS.
(pp. 923 - 931).
ELSEVIER SCIENCE BV
Stereo algorithms can provide useful three-dimensional information for a variety of purposes, including face recognition. For good accuracy, the distance between cameras should be large. However, the large baseline often makes the correspondence problem more difficult. In this paper, we examine the problem of non-Lambertian reflectivity in multi-baseline stereo and propose a refined multiple-baseline stereo technique which estimates the brightness on each camera by using adjacent cameras. In this way, the erroneous stereo correspondences due to brightness changes caused by non-Lambertian reflectance can be reduced without needing to explicitly model the actual reflectance model. Experimental results are presented and significant reductions in correspondences errors are demonstrated in comparison with previous methods. (C) 1997 Elsevier Science B.V.
|Title:||A multiple-baseline stereo for precise human face acquisition|
|Event:||1st International Conference on Audio-Based and Video-Based Biometric Person Authentication (AVBPA 97)|
|Location:||CRANS MONTANA, SWITZERLAND|
|Dates:||1997-03-12 - 1997-03-14|
|Keywords:||multiple-baseline stereo, non-Lambertian reflectivity, Baysian estimation, human face|
|UCL classification:||UCL > School of BEAMS > Faculty of Engineering Science > Computer Science|
Archive Staff Only