AI Infant Posture Detection for Limb Movement Capture
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video surveillance systems fail to accurately and automatically capture dynamic movements of infant limbs during developmental stages like rolling over, sitting, crawling, standing, and walking, focusing instead on facial expressions, which results in inefficient processing of precious images.
Innovation Solution
An image processing method using artificial intelligence to identify and capture target postures of a preset object, such as a baby, by detecting transformations between specific postures like lying supine, lateral recumbent, sitting, crawling, standing, and embracing, and uploading the transformation videos to the cloud for storage when duration thresholds are met.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional video surveillance systems focus on facial expressions for image capture, then facial recognition accuracy is improved, but the ability to capture dynamic limb movements and body posture transformations deteriorates
Solution Approach 1:
The system segments the human body into multiple detection zones including facial region and limb regions. Different detection algorithms are applied to different segments: facial recognition for the face region and posture transformation detection for limb regions, allowing simultaneous optimization of both functions without mutual interference
Solution Approach 2:
The system transitions from two-dimensional facial expression analysis to three-dimensional body posture analysis by incorporating depth information and spatial coordinates of multiple body parts. This dimensional expansion enables comprehensive capture of both facial expressions and limb movements in the same surveillance system
2Measurement precision
If manual selection is used for capturing precious images of infant movements, then image selection accuracy is improved, but processing time efficiency deteriorates
Solution Approach 1:
The system performs preliminary classification of video frames by detecting key posture transformation points in real-time during video recording. When predefined posture transformation conditions are met (e.g., transition from lying to sitting), the system automatically marks and saves relevant frames, eliminating the need for manual review and significantly reducing processing time while maintaining high accuracy
Solution Approach 2:
The system implements real-time feedback mechanisms where detected posture transformations immediately trigger automatic image capture and notification to parents. This closed-loop feedback ensures precious moments are captured without delay while reducing manual intervention to minimal verification steps
3Device complexity
If conventional systems analyze only specific facial expressions, then processing complexity is reduced, but the ability to identify continuous dynamic variations in body posture deteriorates
Solution Approach 1:
The system employs a universal AI processing platform that handles multiple functions: facial expression recognition, body posture detection, limb movement tracking, and milestone identification. This multi-functional architecture maintains manageable complexity through shared computational resources while achieving comprehensive adaptability across different infant development stages and movement types
Data Source
AI summary
An image processing method for human body posture transformation, which is executed by an electronic device reading an executable code to identify a preset object using artificial intelligence, and performing image processing to capture target postures of the preset object. The method includes the steps of identifying an object, detecting postures, and capturing target postures. The steps involve detecting the preset object undergoing transformation between different postures within a target duration. A capture requirement is met when each posture is visible for a posture visibleness duration and reaches a duration threshold. A target posture transformation video which lasts for a segment duration is captured from an initial image and uploaded to the cloud for storage. An electronic device for human body posture transformation image processing, a terminal device in communication connection with the electronic device, and a non-transient computer-readable recording medium are further provided.