User Position Detection via Body Distribution and Face Occlusion Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image detection methods for determining user positions, such as those for monitoring babies, children, patients, or elderly individuals, face challenges with low detection rates due to unstable facial features and discomfort or inefficiency from wearable devices.
Innovation Solution
An image detection method utilizing an artificial intelligence neural network for dual analysis of body distribution and face occlusion, which involves obtaining feature parameters from images, performing region-based analyses, and adjusting confidence levels to accurately determine user positions, including sleeping positions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If facial features are used to determine user position, then detection can be performed without wearable devices, but detection rate is low when user is in lateral face or prone position
Solution Approach 1:
The patent segments the face detection task into multiple view-specific detection models (frontal face, lateral face left, lateral face right, prone position). Each model is specialized for detecting facial features from a particular viewing angle, allowing the system to maintain high detection rates across all positions by selecting the appropriate model based on the detected face orientation.
Solution Approach 2:
The system changes the detection parameters by switching between different detection models based on the user's face position. When the user's face orientation changes (e.g., from frontal to lateral), the system adjusts by selecting a different pre-trained detection model that is optimized for that specific angle, thereby maintaining high detection accuracy across varying positions.
2Reliability
If smart wearable devices are used to obtain physiological information, then continuous monitoring is possible, but power consumption is high and user comfort is reduced
Solution Approach 1:
The patent replaces the mechanical/wearable sensing system with an optical/image-based detection system. Instead of using sensors embedded in wearable devices to detect physiological parameters and position, the system uses image capture devices (cameras) to obtain visual information and derive position and physiological state from image analysis, thereby eliminating the need for powered wearable components.
Solution Approach 2:
The system creates a visual copy or representation of the user's state through images. Rather than directly measuring physiological parameters through contact sensors, the system captures optical copies (images) of the user and extracts positional and physiological information from these visual representations, enabling monitoring without physical contact or wearable power consumption.
3Device complexity
If single analysis method is used for position determination, then processing is simpler, but recognition accuracy is insufficient when confidence level is low
Solution Approach 1:
The system implements a feedback mechanism where the confidence level of the initial position determination is continuously monitored. When the confidence level falls below a predetermined threshold, the system triggers a secondary verification process using face occlusion analysis to re-evaluate and correct the position determination, ensuring high accuracy while maintaining simplicity for clear cases.
Solution Approach 2:
The system applies partial analysis in most cases (simple body distribution analysis when confidence is high) and excessive or additional analysis only when necessary (face occlusion analysis when confidence is low). This selective application of analysis depth optimizes the balance between processing complexity and measurement precision by avoiding unnecessary computational overhead for straightforward cases.
Data Source
AI summary
An image detection method for determining positions of a user. According to the image detection method, a plurality of images of the user are obtained, whether the user moves is determined according to the images, a plurality of feature parameters of the plurality of images are obtained, and a body distribution analysis and a face occlusion analysis are performed to determine the position of the user.