Dynamic Reference Point Selection for Facial Expression Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing facial expression recognition technologies face challenges in accurately recognizing expressions when features points are obscured by items like sunglasses or surgical masks, and are not optimized for varying face orientations and illumination conditions, leading to inaccurate feature extraction and expression recognition.
Innovation Solution
An expression recognition device that acquires images, extracts face areas, determines face conditions based on feature point reliability, and dynamically selects reference points for feature extraction to ensure accurate recognition, regardless of obstructions or environmental factors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If feature points are extracted from images where items like sunglasses or surgical masks are worn, then expression recognition can be performed, but the extraction accuracy of feature points deteriorates
Solution Approach 1:
The system dynamically changes the reference point selection based on detected face conditions (such as presence of sunglasses or surgical masks). When obstructions are detected, the system switches to alternative reference points that are not obscured, thereby maintaining feature extraction accuracy despite the presence of items on the face
Solution Approach 2:
The reference point for feature extraction is made dynamic rather than fixed. The system adapts the reference point selection based on real-time detection of face conditions, allowing the extraction process to respond to varying obstruction scenarios and maintain accuracy across different wearing conditions
2Measurement precision
If a fixed reference point is used for feature extraction, then the extraction process is simple, but accurate expression recognition cannot be performed when facial expressions change
Solution Approach 1:
The reference point selection is made dynamic based on face conditions rather than being fixed. The system automatically adjusts which reference point to use depending on factors such as presence of obstructions, face orientation, and illumination conditions, enabling accurate expression recognition across varying expressions without manual intervention
Solution Approach 2:
The system incorporates feedback mechanisms where the detected face condition (including presence of items, orientation, and illumination) influences the selection of reference points. This feedback loop ensures that the most appropriate reference points are chosen for each specific scenario, maintaining extraction accuracy while adapting to different expression states
3Measurement precision
If feature extraction considers only whether items are worn, then the process is simple, but extraction accuracy deteriorates under varying face orientations and illumination conditions
Solution Approach 1:
The system changes multiple parameters related to face conditions including presence of items, face orientation angles, and illumination characteristics. By considering and adapting to these varied parameters, the system maintains accurate feature extraction across diverse conditions such as different head positions, lighting scenarios, and wearable items
Solution Approach 2:
The reference point selection mechanism is designed to be universal, handling multiple types of face conditions simultaneously. A single adaptive system manages various scenarios including different items worn, multiple face orientations, and varying illumination conditions, making the extraction process robust across all these dimensions without requiring separate specialized processes
Data Source
AI summary
An expression recognition device includes processing circuitry to acquire an image; extract a face area of a person from the acquired image and obtaining a face image added with information of the face area; extract one or more face feature points on a basis of the face image; determine a face condition representing a state of a face in the face image depending on reliability of each of the extracted face feature points; determine a reference point for extraction of a feature amount used for expression recognition from among the extracted face feature points depending on the determined face condition; extract the feature amount on a basis of the determined reference point; recognize a facial expression of the person in the face image using the extracted feature amount; and output information related to a recognition result of the facial expression of the person in the face image.