Switchable Dense Motion Vector Field Interpolation

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention introduces a technique for efficiently exchanging and reconstructing dense motion vector fields in video encoding and decoding. Instead of sending every motion vector, only a sparse set is transmitted, and the missing vectors are recovered by interpolation. What's new is that, for different parts of the image (segments), the optimal interpolation method can be selected and switched between, such as nearest neighbor or bilinear interpolation, improving prediction quality while keeping bitrates low. The segment definitions are derived in a way that the encoder and decoder can reproduce without extra signaling overhead.

Use CasesContent extracted from patent full text and abstract with AI.

• Video codecs for streaming services or video conferencing, enabling higher compression efficiency and better visual quality.
• Applications in mobile video recording and playback, where network or storage resources are limited but high-quality video is desired.
• Real-time video analytics or computer vision systems that benefit from precise motion modeling, such as autonomous vehicles or surveillance.
• Transmission of video over bandwidth-constrained channels, like satellite or rural internet connections.
• Enhancement of virtual reality (VR) and augmented reality (AR) experiences requiring efficient motion compensation for immersive content.

BenefitsContent extracted from patent full text and abstract with AI.

• Reduces the amount of data (bitrate) required to transmit high-quality motion compensation information, improving compression efficiency.
• Enables smoother, more accurate motion field reconstruction, minimizing visual artifacts (such as blockiness) compared to traditional block-based motion compensation.
• Flexible, content-adaptive interpolation per segment ensures optimal quality across diverse motion scenarios within a video frame.
• Segment derivation does not require transmitting additional segmentation information, saving bandwidth and complexity.
• Supports scalability and progressive transmission—denser motion vector information can be added progressively to improve quality as resources allow.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The present disclosure relates to encoding and decoding of motion information. In order to exchange motion information between encoder and decoder, a sparse motion field is provided. In order to improve the reconstruction of the dense motion field, side information indicating a suitable interpolation approach for interpolating the motion vectors of the dense motion field different from the signaled sparse motion field motion vectors is provided. Such information is provided on a segment basis. The segments are derived at the encoder side and the decoder side similarly, based on the reference images and the sparse motion field.