Article

Article title STEREOPAIR DISPARITY ESTIMATE METHOD
Authors V.Ph. Guzik, A.V. Chumachenko
Section SECTION IV. METHODS OF THE ARTIFICIAL INTELLECT
Month, Year 05, 2014 @en
Index UDC 004.923
DOI
Abstract This paper presents stereopair disparity estimate method. Estimate is provided for the whole image as well as for separate image fragments. This allows optimizing pixel correspondence search within a fragment by narrowing the range of possible disparity values. Moreover, splitting the initial image on fragments and calculating mathematical statistics functions on them allows estimating correlation-like approaches robustness within fragments. So, the suggested method allows to mark and handle in some special way textureless image parts on which correlation-like approaches are usually weak. Experiments demonstrate the effectiveness of the suggested method. In particular the expected value of disparity range within fragments is four times less than within the whole stereopair. For some stereopairs disparity lower and upper bounds estimate accuracy reaches 97 %.

Download PDF

Keywords Stereo pair; pixel correspondence; correlation; disparity; statistic function.
References 1. Xing Mei, Xun Sun, Mingcai Zhou, Shaohui Jiao, Haitao Wang, Xiaopeng Zhang. On building an accurate stereo matching system on graphics hardware // Computer Vision Workshops, 2011 IEEE International Conference on Stereo Vision, date of Conference: 6-13 Nov. 2011. – P. 467-474.
2. Klaus A., Sormann M. and Karner K. Segment-based stereo matching using belief propagation and a self-adapting dissimilarity measure // In International Conference on Pattern Recognition (ICPR 2006). – P. 15-18.
3. Z. Wang and Z. Zheng. A region based stereo matching algorithm using cooperative optimization // Computer Vision and Pattern Recognition. – 2008. – P. 1-8.
4. Форсайт Девид А., Понс Жан. Компьютерное зрение. Современный подход: Пер. с англ. – М., 2004. – 928 с.
5. Гузик В.Ф., Чумаченко А.В. Статистический метод оптимизации локальных алгоритмов установления пиксельных соответствий на стереопарах // Известия высших учебных заведений Северо-Кавказский регион. Технические науки. – 2011. – № 4. – С. 20-25.
6. Scharstein D.and Pal C. Learning conditional random fields for stereo // In IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007), Minneapolis, MN, June 2007. URL: http://vision.middlebury.edu/stereo/data (дата обращения 2013.12.30).
7. Hirschmьller H. and Scharstein D. Evaluation of cost functions for stereo matching // In IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007), Minneapolis, MN, June 2007. URL: http://vision.middlebury.edu/stereo/data (дата обращения 2013.12.30).

Comments are closed.