Video Target Tracking with Delayed Key-Frame Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile terminals with limited computing resources experience freezing and reduced user experience due to time-consuming target detection and tracking processes, particularly on key frames, leading to video playback lag.
Innovation Solution
A method that distributes the time-consuming target detection process across multiple delay frames by controlling the detection network to operate during a specified number of delay frames following a key frame, using a tracking network to maintain smooth video playback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If target detection is performed on every video frame, then detection accuracy is improved, but video playback smoothness deteriorates due to time-consuming processing
Solution Approach 1:
The patent segments video frames into key frames and non-key frames, performing target detection only on key frames while using tracking networks on non-key frames. This segmentation reduces the frequency of computationally intensive detection operations while maintaining overall system performance.
Solution Approach 2:
The patent implements periodic target detection by performing detection operations at regular intervals (on key frames) rather than continuously on every frame. This periodic approach balances computational load with detection accuracy requirements.
2Measurement precision
If target detection is performed on key frames, then detection accuracy is improved, but processing time increases causing video lag
Solution Approach 1:
The patent performs preliminary target detection on key frames before video playback, allowing the system to prepare detection results in advance. This preliminary action ensures that when key frames are displayed, the detection results are already available, avoiding playback delays.
Solution Approach 2:
The patent maintains continuous target tracking between key frames using a tracking network, ensuring that target information remains updated and useful throughout the entire video sequence without requiring repeated full detection operations.
3Measurement precision
If detection network operates on every frame, then target tracking accuracy is improved, but device performance deteriorates due to limited computing resources
Solution Approach 1:
The patent divides the processing task into two segments: a detection network that runs on key frames to identify targets, and a tracking network that runs on all frames to maintain target information. This segmentation allows the computationally intensive detection operation to occur less frequently, improving device performance.
Solution Approach 2:
The patent applies partial detection action by performing full target detection only on key frames rather than on every frame. This partial approach provides sufficient detection accuracy while significantly reducing computational burden on devices with limited resources.
Data Source
AI summary
A method for detecting and tracking a target is provided. The method includes: acquiring a target tracking result of each video frame of a plurality of video frames received from a video stream by inputting the each video frame of the plurality of video frames into a tracking network; in response to a video frame of the plurality of video frames being a key frame, acquiring a target detection result output by a detection network upon receiving a last delay frame in a specified number of delay frames by inputting the key frame into the detection network and controlling the detection network to operate during a period of receiving the specified number of delay frames; and generating a final target tracking result based on the target detection result and a target tracking result of the last delay frame.