Abstract—In order to create a good 3D model reconstruction from an image, the shape information of the object of interest is needed. One such information that yields the shape information of an object from images is the shading information. This information can further be utilized using the Shape from Shading method. Unfortunately this method has some drawbacks, namely the depth ambiguity problems. Two points in the image which have the same intensity will have the same depth, although in reality they have different depth. Color also plays a huge effect to this ambiguity. Some color generates higher intensity than the others. This makes the system generate an incorrect reconstruction of the image. We propose a color segmentation based depth adjustment to overcome this problem. First we cluster the image into color clusters. Based on this color cluster we will adjust the depth of vertices belonging to the higher intensity cluster accordingly. The proposed method can reduce the effect of these depth ambiguity problems.
Index Terms—3D model reconstruction, shape from shading, color segmentation.
The authors are with the New Industry Hatchery Center (NICHe) of Tohoku University, Sendai, Miyagi Prefecture 9808579, Japan (e-mail: vicky@ riec.tohoku.ac.jp, aoki@riec.tohoku.ac.jp).
[PDF]
Cite:Vicky Sintunata and Terumasa Aoki, "Color Segmentation Based Depth Adjustment for 3D Model Reconstruction from a Single Input Image," International Journal of Computer Theory and Engineering vol. 8, no. 2, pp. 171-176, 2016.