Always-On Camera User Identification Through Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic devices struggle to continuously obtain personalized user information without direct user input and accurately distinguish user data from non-user data, leading to noise in user input-based applications.
Innovation Solution
An always-on camera and sensor system that continuously captures image and sensor data, processes features, performs clustering, and generates person identifiers to determine device idle states, storing relevant data for personalized services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user input-based information acquisition is used, then information about user can be obtained accurately, but continuous information acquisition is not possible and device must be actively used by user
Solution Approach 1:
The system performs preliminary actions by continuously capturing images and sensor data in advance before user interaction is needed. The always-on camera and sensor system collect data proactively, so when user information is later needed, the data is already available for processing and analysis.
Solution Approach 2:
The device performs self-service by automatically collecting, processing, and analyzing user information without requiring active user participation. The system autonomously monitors the environment, extracts features from images and sensor data, and generates person identifiers independently of user input.
2Productivity
If sensor data analysis is performed without user input, then continuous monitoring is possible, but it is difficult to directly relate sensor data to specific user
Solution Approach 1:
The system introduces an intermediary approach by combining multiple data sources (images from always-on camera, sensor data, and extracted features) to bridge the gap between continuous monitoring and specific user identification. The feature extraction and clustering processes act as intermediaries that transform raw data into identifiable user profiles.
Solution Approach 2:
The system changes parameters by transforming raw sensor data and image data into extracted features through processing. By converting data from its original form into feature representations and then into person identifiers through clustering, the system achieves both continuous monitoring and accurate user identification.
3Productivity
If always-on camera and sensor data are collected continuously, then continuous user information acquisition is enabled, but data processing complexity and storage requirements increase
Solution Approach 1:
The system extracts only the necessary information from continuous data streams by identifying and isolating relevant features from images and sensor data. Through feature extraction, the system separates meaningful user-related information from unnecessary data, reducing processing complexity while maintaining continuous monitoring capability.
Solution Approach 2:
The system applies local quality by processing and storing only the most relevant data characteristics. Through feature extraction and clustering, the system focuses computational resources on identifying and storing meaningful patterns (person identifiers) rather than processing all raw data uniformly, thereby managing complexity.
4Measurement precision
If clustering is performed on extracted features, then user identification accuracy is improved, but processing time and computational resources increase
Solution Approach 1:
The system performs preliminary feature extraction and prepares data for clustering in advance. By pre-processing images and sensor data into features and organizing them for clustering, the system reduces the computational burden during actual user identification, thereby decreasing processing time while maintaining accuracy.
Data Source
AI summary
According to the present disclosure, a method performed by an electronic device may include: obtaining image data using an always-on camera, obtaining sensor data, obtaining combined data from the image data and the sensor data based on an obtained time of the image data and an obtained time of the sensor data, extracting at least one feature based on the combined data, generating and storing at least one feature set based on the at least one feature, performing clustering on the at least one feature set, and storing a result of performing the clustering.