Acquiring/Recognising Facial Features TRIZ Cases
- HMD Facial Expression Synthesis via Mouth Image Analysis
A system estimates user expressions from mouth images to produce synthetic facial avatars for live broadcasts.
- Landmark-Free Face Attribute Prediction via Learned Spatial Transformations
Learned spatial transformations align unconstrained faces for attribute prediction, bypassing landmark detection errors from pose and occlusion.
- Facial Stroke Detection Using Feature Symmetry and Classifier
A facial stroke detection system extracts image features to form a determining set for a classifier.
- Facial Feature Tracking via Bidirectional Temporal Graph Optimization
Bidirectional temporal graph optimization refines facial feature location hypotheses across video frames for robust tracking.
- Multimedia Deception Detection via Multi-Tier Classification
Multi-tier classification analyzes facial, vocal, and textual features to score deception likelihood in multimedia content.
- Facial Expression Determination Using Angle Estimation
Segmenting facial regions and estimating orientation angles maintains expression determination accuracy for tilted faces.
- Vehicle Emotion Recognition Service Provision via Temporal Pattern Analysis
A vehicle emotion recognition system classifies occupant feelings into transient, sequential, or repetitive forms using temporal analysis.
- Facial Expression Library Creation for Engaging Online Learning
Image processing identifies facial landmarks to generate an animated avatar library from speaker video.
- Object Emotion Analysis Model Fusing Static and Dynamic Features
A pre-trained object emotion analysis model fuses static facial features with dynamic expression, sound, and language content data.
- Audience Interest Evaluation Using Smile and Face Orientation Detection
A detection system combines smile intensity with face orientation metrics to assess audience attention levels.
- Face Liveness Detection via Multi-View Image Variance Analysis
A face liveness detection system verifies human presence by comparing facial appearance variance across multiple captured images.
- Normalized Virtual Camera for Facial Feature Detection
A normalized virtual camera generates perspective-projected images to preserve geometric features for AI facial analysis.
- Sensor Data Analysis System for Real-Time Customer Profile Synthesis
Machine learning system processes video and audio sensor data to construct detailed customer profiles for personalized user experiences.
- Imaging Apparatus Subject Priority Control via Date Association
A control unit prioritizes image capture for subjects linked to the current date, aligning automatic operation with user intent.
- Depth Tensor Train Network for Classroom Behavior Detection
Tensor train decomposition reduces hardware requirements while maintaining image recognition accuracy for embedded teaching assistance systems.
- Self-adaptive AI for Real-time Video Surveillance
A self-adaptive AI system generates space-time-behaviour datasets from video frames to identify events and causes in real time.