Image Presentation Timing With EEG Feedback for Missed Detection Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image recognition technologies face challenges such as unbalanced training data distribution, noise in training data, difficulty in extracting high-order semantic features, and high miss detection rates due to brain fatigue and limited attention resources in brain-computer collaboration systems.
Innovation Solution
A method and device that adjust the presentation time of image sequences based on a duration impact parameter and fatigue state parameter to optimize brain-computer collaboration, combining computer vision algorithms with electroencephalogram feedback signals to reduce miss detection rates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image recognition processing is performed on all captured images, then recognition accuracy is improved, but processing time and power consumption increase
Solution Approach 1:
The patent applies partial action by performing full image recognition processing only on selected images that meet specific criteria (such as containing potential falling objects), while using simplified processing or no processing on other images. This selective approach maintains recognition accuracy for critical cases while significantly reducing overall processing time and computational resources.
Solution Approach 2:
The patent segments the image processing workflow into multiple stages: initial filtering to identify candidate images, followed by detailed recognition processing only on those candidates. This segmentation allows the system to maintain high accuracy for important detections while minimizing processing overhead by excluding irrelevant images from intensive analysis.
2Measurement precision
If image recognition processing is performed on all captured images, then recognition accuracy is improved, but power consumption increases
Solution Approach 1:
The patent implements partial processing by applying full recognition algorithms only to images that pass initial screening criteria, thereby maintaining recognition accuracy for critical detections while substantially reducing power consumption by avoiding processing of unnecessary images.
Solution Approach 2:
The system uses automatic image selection based on predefined criteria (such as detecting potential falling objects in preview images) to determine which images require full recognition processing. This self-service mechanism eliminates the need for continuous full-processing of all captured images, thereby reducing power consumption while maintaining accuracy for important cases.
3Ease of operation
If presentation time is fixed, then device operation is simple, but image recognition accuracy deteriorates due to insufficient processing time
Solution Approach 1:
The patent implements dynamic presentation time adjustment where the display duration of captured images is automatically adapted based on recognition results and image characteristics. This dynamic approach allows the system to maintain simple operation for users while ensuring sufficient processing time for accurate recognition by extending display time only when necessary for critical detections.
Solution Approach 2:
The system changes the presentation time parameter dynamically based on recognition outcomes and image content characteristics. By adjusting this parameter rather than using a fixed value, the system achieves both simple operation and high recognition accuracy, as the presentation time is optimized for each specific case without requiring complex user intervention.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments of this application provide an image recognition method and device, and an image presentation time adjustment method and device. The image recognition method includes: setting a presentation time sequence corresponding to an image sequence, where the image sequence includes N images, the presentation time sequence includes at least two unequal presentation times, a difference between any two presentation times of the at least two unequal presentation times is k x Δ, k is a positive integer, and Δ is a preset time period value; processing the image sequence by using a computer vision algorithm, to obtain a computer vision signal corresponding to each image in the image sequence; obtaining a feedback signal that is corresponding to each image in the image sequence and that is generated when an observation object watches the image sequence displayed in the presentation time sequence; and fusing, for each image in the image sequence, a corresponding computer vision signal and a corresponding feedback signal to obtain a target recognition signal of each image in the image sequence. Implementation of the embodiments of this application can reduce a miss detection rate of image recognition.