Multimedia Choking Detection via Mouth Gesture Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems fail to monitor and prevent choking incidents in elderly and children while they are engrossed in multimedia consumption, leading to fatal mishaps due to lack of real-time monitoring of chewing and swallowing patterns.
Innovation Solution
A system and method that utilize an image capturing module, mouth gesture identification module, training module, and prediction module to capture and analyze facial features and mouth gestures to build personalized support models, pausing multimedia and alerting users or caregivers when choking is detected.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multimedia is played continuously without interruption, then user entertainment experience is improved, but user safety deteriorates due to undetected choking incidents
Solution Approach 1:
The system performs preliminary action by capturing images and analyzing mouth gestures before a choking incident becomes critical. The image capturing module continuously monitors the user's mouth area, and the analysis module detects swallowing patterns in advance, allowing the system to pause multimedia proactively before a choking event occurs, thus preventing harm while maintaining relatively simple system architecture.
Solution Approach 2:
The patent replaces complex mechanical monitoring systems with an optical-based image analysis system. Instead of using mechanical sensors or contact-based detection devices, the system uses an image capturing module to take pictures and an analysis module to process visual data of mouth gestures, substituting mechanical complexity with optical and computational approaches that are less intrusive and easier to integrate.
2Measurement precision
If image capturing frequency is increased to improve detection accuracy, then choking detection precision is improved, but energy consumption increases
Solution Approach 1:
The system implements periodic action by capturing images at specific intervals rather than continuously. The image capturing module takes pictures at predetermined time intervals during multimedia playback, and the analysis module processes these periodic images to detect swallowing patterns. This periodic sampling maintains adequate detection precision while significantly reducing energy consumption compared to continuous high-frequency capture.
Solution Approach 2:
The system applies partial action by focusing image capture and analysis only on critical moments when swallowing is expected to occur during multimedia consumption. Rather than monitoring all activities equally, the system activates enhanced monitoring specifically during eating periods detected through initial analysis, using partial computational resources to achieve high precision where needed while conserving energy during non-critical periods.
3Adaptability or versatility
If personalized support model is built for each user, then system adaptability is improved, but data processing time increases
Solution Approach 1:
The system performs preliminary action by building personalized support models during an initial training phase before actual use. The training module collects user-specific data during a calibration period and generates individualized swallowing pattern models in advance. Once built, these pre-personalized models are stored and reused during normal operation, enabling high adaptability to each user without incurring processing delays during critical monitoring periods.
Data Source
AI summary
A system for controlling viewing of multimedia is provided. The system includes an image capturing module captures images or videos of a user while viewing the multimedia. A mouth gesture identification module extracts facial features from the images captured of the user; identifies mouth gestures of the user based on the facial features extracted. A training module analyses the mouth gestures identified to determine parameters; builds a personalised support model for the user based on the parameters determined. A prediction module receives real-time images captured, wherein the real-time images are captured while viewing the multimedia; extract real-time facial features from the real-time images captured; identifies real-time mouth gestures of the user based on the real-time facial features extracted; analyze the real-time mouth gestures identified to determine real-time parameters; compare the real-time parameters determined with the personalized support model built for the user, and control outputs based on compared data.