Encoded Video Object Detection Using Metadata-Guided Frame Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video stream processing methods are resource-intensive as they uniformly process and decode all frames, including those without objects of interest, leading to unnecessary resource consumption.
Innovation Solution
An object detection resource gateway analyzes metadata of non-key frames to identify frames with objects of interest, selectively decoding and processing only those frames, thereby reducing resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all video frames are continuously processed and transmitted for object detection, then detection accuracy is maintained, but network and processing resources are wasted
Solution Approach 1:
The patent applies preliminary action by analyzing metadata (such as motion vectors and packet sizes) before performing full frame decoding and object detection. This preliminary analysis identifies frames containing objects of interest, allowing the system to skip unnecessary processing of empty frames while maintaining detection accuracy for relevant frames.
Solution Approach 2:
The patent extracts and analyzes specific metadata elements (motion vectors, packet sizes, frame differences) from the video stream without processing the entire frame data. This extraction approach enables resource-efficient identification of frames containing objects of interest, separating the detection task from full frame processing.
2Loss of energy
If metadata analysis is used to skip frame decoding, then resource usage is reduced, but detection accuracy may be compromised
Solution Approach 1:
The patent implements feedback by using detected objects from previously decoded frames to inform subsequent frame processing decisions. When objects are detected, the system adjusts its metadata analysis thresholds and continues decoding subsequent frames, creating a feedback loop that maintains detection accuracy while optimizing resource usage based on actual detection results.
Solution Approach 2:
The patent changes parameters dynamically by adjusting detection thresholds and processing intensity based on metadata analysis results. When motion vectors or packet sizes indicate potential objects of interest, the system increases processing intensity; when metadata suggests empty frames, it reduces processing, thereby maintaining accuracy while optimizing resource consumption.
Data Source
AI summary
The description relates to resource aware object detection for encoded video streams that can identify frames of the video stream that include an object of interest, such as a human, without decoding the frames.