Motion Metadata Derivation for Encoded Video Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining motion metadata in recorded videos are either unavailable due to camera limitations or require processor-intensive frame-by-frame comparisons, especially when search parameters change, which is time-consuming and power-demanding, especially when dealing with encoded video formats.
Innovation Solution
A method that decodes encoded video frames, divides them into a mesh of cells, calculates a metric of change by comparing pixel data across frames, and stores this data as motion metadata, allowing for efficient identification of motion within specified regions of interest using pre-computed thresholds, reducing the need for repeated frame-by-frame comparisons.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If frame by frame comparison is performed every time search parameters change, then motion events can be identified, but processing power consumption increases and time is consumed
Solution Approach 1:
The patent pre-computes motion metadata by performing frame-by-frame comparison once during video encoding or preprocessing, storing the results as motion metadata. When search operations are needed, the system directly queries the pre-computed motion metadata instead of re-performing frame comparisons, significantly reducing processing time for motion event identification.
Solution Approach 2:
The patent creates a copy of motion information in the form of motion metadata that mirrors the essential motion characteristics extracted from frame comparisons. This copy allows multiple search operations to be performed on the stored metadata without needing to access or re-process the original video frames, enabling efficient repeated queries with different search parameters.
2Measurement precision
If frame by frame comparison is performed on encoded video, then pixel changes can be detected, but decoding is required which further adds processing power usage and time
Solution Approach 1:
The patent performs the computationally intensive frame-by-frame comparison and decoding operations in advance during video preprocessing or encoding, generating motion metadata that captures pixel change information. This preliminary action eliminates the need to decode and re-compare frames during subsequent search operations, significantly reducing processing power consumption when identifying motion events.
Solution Approach 2:
The patent creates a copy of motion information in the form of motion metadata that preserves the essential pixel change characteristics without requiring access to the original encoded video data. This metadata copy allows motion detection operations to be performed on compressed numerical data rather than decoded video frames, reducing processing power requirements.
3Ease of operation
If camera collects metadata about moving objects, then motion information is available, but camera must support collection and user must enable it
Solution Approach 1:
The patent creates a universal copy of motion information in the form of standardized motion metadata that can be generated from any video source regardless of camera capabilities. This approach eliminates dependency on camera-specific motion detection features, allowing motion metadata to be created from standard video encodings from any camera or video source.
Solution Approach 2:
The patent introduces motion metadata as an intermediary representation that bridges the gap between raw video data and motion analysis requirements. This intermediary layer provides a standardized interface for motion information that is independent of the original video source, enabling consistent motion event identification across different cameras and video formats.
Data Source
AI summary
A video processing device which generates motion metadata for encoded video comprises a decoder configured to decode frames of an encoded video into image frames; and a processing circuitry configured to execute a motion meta data deriving operation on image frames decoded by the decoder. The motion meta data deriving operation comprises: a dividing function configured to divide a current image frame into a mesh of cells, wherein each cell comprises multiple image pixels, a comparison function configured to determine a metric of change for each cell by comparing pixel data of each cell with pixel data of a correspondingly positioned cell of a previous and/or subsequent image frame, and a storing function configured to store the metric of change for each cell as the motion metadata related to the current image frame.