A way to generate new video frames
A video frame and video technology, applied in the field of video editing, can solve problems such as inaccurate estimation, poor effect of new video frames, and low efficiency
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment Construction
[0062] For a video clip with a limited number of video frames, you can use figure 1 The method described is implemented.
[0063] The specific implementation method is:
[0064] First, two frames before and after the video (I 1 , I 2 ) input optical flow estimation network PWCNET [6] In , pass through a dilated convolutional layer and calculate the correlation of the generated features; finally, input the cv into 5 convolutional layers with a size of 3×3 and feature numbers of 128, 128, 96, 64, and 32 to generate light flow (f 1→2 H / 4,W / 4 , f 2→1 H / 4,W / 4 ); Use bilinear interpolation to downsample the original video (I 1 H / 4,W / 4 , I 2 H / 4,W / 4 ), and then combine it with the optical flow (f 1→2 H / 4,W / 4 , f 2→1 H / 4,W / 4 ) cascaded into the first sub-network NET 1 , learn the optical flow between the new video frame and the two frames before and after (f t→1 H / 4,W / 4 , f t→2 H / 4,W / 4 ) and its corresponding mask (m 1 H / 4,W / 4 ,m 2 H / 4,W / 4 );
[0065] Next, ...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com