Two-Level Wakeup Detection for Accurate Human Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing smart door locks struggle with low recognition accuracy in distinguishing between human and animal movements, leading to frequent false, missed, or delayed alarms, which undermines user experience and reliability.
Innovation Solution
A two-level wakeup network system is employed, where a level-1 wakeup network performs initial detection, followed by a higher precision level-2 network, to accurately identify human figures in a single frame, reducing false and missed alarms while minimizing power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single-level detection network is used, then power consumption is reduced, but recognition accuracy deteriorates
Solution Approach 1:
The detection system is segmented into two levels: a level-1 wakeup network for initial detection and a level-2 wakeup network for refined detection. The level-1 network processes all incoming images to identify potential targets, and only images with positive detections are forwarded to the level-2 network. This segmentation allows the system to maintain high recognition accuracy through the level-2 network while the level-1 network filters out false positives, reducing overall power consumption by avoiding unnecessary processing of all images at full resolution.
2Measurement precision
If two-level wakeup network is implemented, then recognition accuracy is improved, but device complexity increases
Solution Approach 1:
The level-1 wakeup network performs preliminary action by conducting initial detection on all incoming images before they are processed by the level-2 network. This preliminary screening identifies potential targets and filters out non-targets, so that the level-2 network only processes images that need detailed analysis. This preliminary action reduces the burden on the level-2 network and simplifies the overall system operation, managing complexity through structured processing stages.
3Measurement precision
If high-frame-rate SMD is triggered frequently, then detection precision is improved, but power consumption increases
Solution Approach 1:
The system dynamically adjusts the detection frame rate based on the detection results from the level-1 network. When the level-1 network detects no potential targets, the system operates at a lower frame rate to save power. When the level-1 network identifies potential targets, the system switches to high-frame-rate SMD mode to ensure precise detection. This dynamic adjustment allows the system to maintain high detection precision when needed while significantly reducing power consumption during periods when high precision is not required.
Data Source
AI summary
A wakeup method includes: obtaining an image; using the image as input data of a level-1 wakeup network to obtain a first detection result output by the level-1 wakeup network, where the level-1 wakeup network is used to perform target object detection on the input image; waking up a level-2 wakeup network when the first detection result indicates that a target object exists in the image, where the level-2 wakeup network is used to perform target object detection on the image, and detection precision of the level-2 wakeup network is higher than detection precision of the level-1 wakeup network; detecting the image by using the level-2 wakeup network to obtain a second detection result output by the level-2 wakeup network; and when the second detection result indicates that the target object exists in the image, waking up a processing unit to perform a preset operation.