Image Processing Trifocal Depth Capture

zeile_1
text_1
While benefits of depth maps for content creation and 3D post-production for multi-view displays are already widely known, their usage as an alternative stereo 3D (S3D) production workflow is just starting to become evident. Researchers of Fraunhofer Heinrich Hertz Institute HHI have developed a system based on Trifocal Depth Capture providing a way towards reliable depth maps with only minimal additional effort on set compared to 2D. During post-production dense depth maps are estimated allowing to render additional virtual views with an individual virtual camera baseline optimally adjusted for the planned screening scenario.
contact_tanger
zeile_2
zwischenü
Technical Background
technical background
- A third camera detects uncertainties of estimated depth maps by allowing cross-checks.
- The declaration of reliability values provides means for guided fusion with other sensors and final filtering.
- A special real-time configuration can be used on set for immediate preview of depth maps in lower resolution.
- An offline high-quality configuration for postproduction finally provides a dense high resolution depth map for the center camera, optimized for rendering of new virtual views.
- Plug-ins for Adobe AfterEffects offer a seamless integration into existing workflows
zwischenü
Benefits
Benefits
- Workflow on set similar to 2D
- Camera setup is fixed
- Preview on set allows measuring without tape
- No dependency on shot depth range
- No dependency on targeted presence factor
- Flexible rendering of second view during post-production
- Future proof for autostereoscopic displays
zwischenü
References
References
- Presented on SMPTE 2nd Annual International Conference on Stereoscopic 3D, June 2011
- This work has been partially funded by the EC, contract FP7 215475 “2020 3D Media” (www.20203dmedia.eu)
infobox
multiview_generation_text
prev_events
infografik
Offline high-quality depth map suitable for rendering of new virtual views
infografik2
Sparse reliable depth map (left) and full depth map (right). Red depicts background parts, green objects in the middle and blue near objects.
