Title :
An improved power cepstrum based stereo correspondence method for textured scenes
Author :
Smith, Philip W. ; Nandhakumar, N.
Author_Institution :
Dept. of Electr. Eng., Virginia Univ., Charlottesville, VA, USA
fDate :
3/1/1996 12:00:00 AM
Abstract :
This paper analyses the performance of cepstral approaches for solving the stereo correspondence problem. A quantitative analysis of the effects of noise, foreshortening differences, and photometric variations on existing cepstral correspondence techniques is presented. A modified approach that is less sensitive to these effects is developed for textured scenes, and analytical arguments for its robustness are developed. The results of a comparative study of the new cepstral technique, the original cepstral algorithm and the cross-correlation approach are shown and discussed. The performance of the new method is experimentally verified on textured surfaces
Keywords :
cepstral analysis; computer vision; filtering theory; image texture; stereo image processing; cross-correlation; depth estimation; foreshortening correction; noise effect; photometric variations; power cepstrum; quantitative analysis; stereo image correspondence; stereopsis; textured scenes; textured surfaces; Cepstral analysis; Cepstrum; Filtering; Image texture analysis; Layout; Noise robustness; Performance analysis; Photometry; Stereo vision; Surface texture;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on