Stereoscopic Window Metadata for Dynamic Edge Mismatch Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Stereoscopic window violations in stereoscopic video content cause viewer discomfort due to mismatched left and right stereo eye views, leading to issues like retinal rivalry, which existing solutions like the 'floating window' technique cannot adjust post-production or during presentation.
Innovation Solution
Generate stereoscopic window violation metadata information to define edge offsets for each image pair, allowing dynamic adjustment of video frames on-the-fly to mitigate these violations, using techniques such as cropping, changing pixel color or transparency, and encoding this metadata with the video content.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the floating window technique is used to crop edge images to match stereo eye views, then stereoscopic window violations are mitigated, but the image source is permanently affected and cannot be adjusted later in post-production or during presentation
Solution Approach 1:
The patent separates the stereoscopic video content from the window violation mitigation data by using independent metadata tracks. Each video frame is associated with corresponding mitigation data that defines cropping parameters, allowing the video content to remain intact while enabling flexible application of window violation mitigation through separate, modifiable metadata tracks that can be adjusted independently without affecting the original video source
Solution Approach 2:
The patent enables dynamic adjustment of window violation mitigation by storing mitigation parameters in modifiable metadata tracks rather than permanently altering the video content. This allows the cropping and masking parameters to be changed dynamically during post-production and presentation, adapting to different convergence settings and viewing conditions while maintaining the original video quality
2Object-affected harmful factors
If edge images are cropped to eliminate monocular objects, then retinal rivalry is reduced, but image quality and content are permanently altered
Solution Approach 1:
The patent introduces metadata tracks as an intermediary layer between the original video content and the display process. This intermediary contains cropping and masking parameters that define how to mitigate window violations without permanently altering the video source. The metadata acts as a flexible instruction set that can be applied dynamically, preserving original image quality while eliminating retinal rivalry through controlled cropping and transparency adjustments
Solution Approach 2:
The patent creates a separate copy of the mitigation parameters in metadata tracks rather than modifying the original video frames. This copying approach allows the same original video content to be used with different mitigation parameters for different scenes or convergence settings, preserving the master copy's quality while enabling flexible mitigation applications
Data Source
AI summary
Techniques are disclosed for addressing “stereoscopic window violations” in stereoscopic multimedia content. Stereoscopic window violations result in the stereoscopic effect becoming “broken” for the viewer and may occur, e.g., when the left and right stereo eye views in the stereoscopic content are mismatched. Stereoscopic mismatch often occurs at the edges of the left and right eye video image frames (wherein, e.g., a depicted shape may become cut off for a left eye view but not a corresponding right eye view). According to the techniques disclosed herein, rather than permanently masking or otherwise editing the stereoscopic video content to account for any window violations, accompanying stereoscopic window violation metadata information may be generated for the stereoscopic video, which may be used to define a particular geometry for each left and right eye video image frame pair, and used at playback time to crop, mask, or otherwise modify the video image frames.