Camera-Based Viewer Immersion Detection Using AI Image Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audience rating survey methods, such as people meters, face challenges in accurately determining viewer immersion due to participant activeness and the risk of privacy invasion when using camera-based image recognition, as they often rely on manual input and cannot distinguish between intense and distracted viewing states effectively.
Innovation Solution
A data generation method and apparatus that utilizes a camera-based image recognition system to extract image recognition data for face, posture, and multitasking identification, applying additional or deduction points based on preset criteria to determine viewer immersion without revealing personal identification information, thereby generating accurate viewing immersion data sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual input method is used for audience rating survey, then participant activeness can be monitored, but measurement precision of viewing immersion is insufficient
Solution Approach 1:
The patent replaces manual input mechanisms with automated image recognition technology. The camera captures viewer images, and AI algorithms automatically analyze facial expressions, posture, and multitasking behaviors to determine viewing immersion, eliminating the need for manual participant input while significantly improving measurement precision.
Solution Approach 2:
The system enables self-service measurement by automatically capturing and analyzing viewer states without requiring active participation or input from viewers. The camera-based system continuously monitors and evaluates viewing immersion independently, making the survey process passive for participants while maintaining high measurement accuracy.
2Measurement precision
If camera-based image recognition is used to extract viewer images, then viewing immersion can be determined, but privacy invasion risk increases
Solution Approach 1:
The patent extracts only the necessary features for viewing immersion assessment from captured images, such as facial expression patterns, posture characteristics, and multitasking indicators. Personal identification information and other sensitive data are deliberately excluded from analysis and processing, maintaining measurement precision while minimizing privacy risks.
Solution Approach 2:
The system introduces an intermediary processing layer that transforms raw image data into abstract viewing state metrics. Instead of directly using or storing personal images, the system processes them through AI algorithms that generate anonymized immersion scores, acting as a mediator between image capture and data storage to protect viewer privacy.
3Productivity
If manual input method is used for survey, then system complexity is low, but productivity of data collection is reduced
Solution Approach 1:
The camera-based system enables continuous automatic data collection without interruption or manual intervention. Unlike manual input methods that require active participant cooperation and result in gaps in data collection, the image recognition system continuously captures and analyzes viewer states, significantly improving data collection productivity and coverage.
Solution Approach 2:
The patent replaces manual data input processes with automated image recognition technology. The system automatically captures images, analyzes viewing states, and generates immersion data without human intervention, eliminating the bottlenecks and inefficiencies of manual survey methods while managing system complexity through modular AI processing.
Data Source
AI summary
Provided is a data generation apparatus including an image reader configured to extract image recognition data by performing image recognition for identifying face, posture, and multitasking of a user viewing an image, and transmit the image recognition data to a score processing unit; the score processing unit configured to apply an additional or deduction point according to a preset conditional value of the additional and deduction point for immersion determination to the image recognition data extracted from the image reader and convert an immersion determination value to a numerical value; an immersion determinator configured to determine an immersion state for each time section based on a preset criterion with respect to the immersion determination value converted by the score processing unit; and a combined data generator configured to combine a user identification (ID), viewing channel information, time information, and an immersion state and generate the same as a data set.