Development of an optical 3D human face reconstruction system using moving-window method

This paper introduces a stereo vision system for rapid and accurate geometry acquisition of human faces. A moving-window algorithm is proposed to enhance the consistency in searching for corresponding locations on images captured from different views. Effectiveness of the approach is demonstrated by a comparison of correctness with two of the most popular algorithms, the Scale Invariant Feature Transform (SIFT) and the Phase-Only Correlation method.