Efficient Automatic Depth Estimation for Video
Supplemental Results [10.1109/ICDSP.2013.6622807]

On this page you'll find the supplemental results for "Efficient Automatic Depth Estimation for Video". These are the full-resolution versions of the results presented in the paper. Also included are the videos of the frame sequences presented at the end of the paper in the results section.

Depth Interpolation

The interpolation section has two sets of results. The first is a comparison between the old interpolation method¹ and the proposed one. The second set is an example of how the depth map can be adjusted by a user.

Comparison

• 
  Interpolation using the original method¹.
• 
  Interpolation using edge-aware filtering.

Map Correction

• 
  The source image with the overlaid user correction scribble.
• 
  The corrected depth map.

Results

The results section contains the full-length generated sequences presented in the paper. Please go to the Consistent Depth Maps Recovery from a Video Sequence website for the original frame sequences and ground truth depth maps.

Note: Your browser must support the HTML5 <video> tag in order to view the videos below.

Road

Flower Pot

References

1. R. Rzeszutek, R. Phan, and D. Androutsos, "Depth estimation for semi-automatic 2d to 3d conversion," in Proceedings of the 20th ACM international conference on Multimedia, ser. MM '12. New York, NY, USA: ACM, 2012, pp. 817-820.