Neutral Face Expression Capture for Vehicle Occupant Emotion Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing emotion estimation technologies face challenges in efficiently acquiring a vehicle occupant's face image with a neutral facial expression, leading to inaccurate emotion estimation due to forced or unnatural expressions, especially in dynamic situations like traveling on curves or highways.
Innovation Solution
An information processing apparatus that automatically captures and stores a vehicle occupant's face image as neutral when the occupant is likely to have a plain expression, using cameras and sensors to monitor vehicle situations and associate the image with identification information, allowing for sharing across vehicles to improve estimation accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If face images are captured continuously to ensure neutral expressions are acquired, then the probability of acquiring neutral expressions improves, but the system complexity and data processing burden increase
Solution Approach 1:
The system pre-registers multiple candidates for neutral expression situations (e.g., vehicle stationary, straight road travel) before actual emotion estimation is needed. When a neutral expression is required, the system checks if the current situation matches any pre-registered neutral situations, allowing rapid retrieval without continuous capture or complex real-time analysis.
Solution Approach 2:
The system uses the vehicle's existing operational data (speed, acceleration, steering angle) to automatically determine neutral expression conditions without requiring additional sensors or complex external validation. The vehicle's own operational state serves as the criterion for identifying neutral expressions.
2Quantity of substance
If face images are captured in all driving situations, then comprehensive emotion data is obtained, but the accuracy of neutral expression identification decreases due to dynamic conditions
Solution Approach 1:
The system applies different capture strategies to different driving situations. Neutral situations (stationary, straight road) use continuous or frequent capture to ensure high-quality neutral images, while non-neutral situations (curves, acceleration) use reduced capture frequency or skip capture entirely, optimizing both quantity and quality locally for each situation type.
Solution Approach 2:
The system pre-identifies and categorizes driving situations into neutral and non-neutral types based on vehicle operational parameters. This preliminary classification allows the system to selectively capture images only in appropriate situations, ensuring high neutral expression identification accuracy while maintaining sufficient image quantity.
3Measurement precision
If multiple face images are stored for each occupant, then the most recent neutral expression can be selected, but the data storage and management complexity increases
Solution Approach 1:
The system pre-organizes face images by associating each image with its corresponding neutral situation type and timestamp. When a neutral expression is needed, the system queries for images from the current situation type, automatically selecting the most recent appropriate image without requiring complex sorting or filtering operations.
Solution Approach 2:
The system creates simplified metadata copies of face image data that include only essential information (situation type, timestamp, identification). This lightweight copying approach allows efficient data management and retrieval without handling the full complexity of the original image data.
Data Source
AI summary
An information processing apparatus comprising a situation acquiring unit configured to acquire a situation of one movable object, and a storing execution unit configured to store a face image of an occupant on the one movable object, captured by an image-capturing unit mounted on the one movable object when the one movable object is in a predetermined situation, as a face image of the occupant with neutral facial expression is provided.