Stereo matching method using color frequency information
Autor: | Toshiyuki Kashiwagi, P. Phonsue, Shunichiro Oe |
---|---|
Rok vydání: | 2006 |
Předmět: |
Color histogram
Computer science Color image business.industry Template matching Feature extraction ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION Pattern recognition Feature (computer vision) Computer Science::Computer Vision and Pattern Recognition Face (geometry) Computer vision Artificial intelligence business Computer stereo vision Feature detection (computer vision) |
Zdroj: | IEEE International Conference Mechatronics and Automation, 2005. |
DOI: | 10.1109/icma.2005.1626543 |
Popis: | Stereo method needs the search of corresponding points to obtain depth information. However, the conventional stereo method cannot apply to the object with uniform or almost uniform surface. So we propose the new stereo matching method applicable to such object. From the characteristic of color histogram, the non-uniform part in such the object has lower frequency value than that of the uniform part. By using this characteristic, the image feature of the non-uniform part can be extracted and the distance measurement is performed by matching for this image feature. In this paper the validity of proposed method is verified by the experiment obtaining the depth information for the object with almost uniformity and human's face. |
Databáze: | OpenAIRE |
Externí odkaz: |