Scene Recognition TRIZ Cases
- Construction Object Detection Using Brightness and Color Channels
A vehicle perception system converts camera images into brightness and color channels to identify potential construction objects using template matching.
- Optical Sensor Region Segmentation for Autonomous Vehicle Object Detection
Information processing device segments optical sensing data using map and position information to target object detection models.
- On-board Camera Exposure Control for Stereo Vision Accuracy
A stereo vision camera system switches exposure conditions to capture accurate three-dimensional data and object recognition images.
- Information Processing Apparatus Selective Data Labeling
An information processing apparatus sets labels on scene data to determine storage eligibility.
- Steerable Landing Light System for Crosswind Runway Illumination
A steerable landing light system uses a camera and controller to aim illumination along the aircraft's actual travel path.
- Modality-Independent Feature Extraction for Multi-Sensor Object Classification
A feature extraction unit generates modality-independent features from sensor data to enable a single classification unit across multiple input types.
- AI Blinking Light Detection Using Image Encoding Sequences
AI models encode image frames to detect blinking vehicle lights, reducing data bandwidth and storage needs.
- Vehicle Radar QPSK Beamforming Sidelobe Reduction
QPSK beamforming reduces sidelobe energy loss while improving vehicle hazard identification.
- Hyperspectral Training Data for AI Remote Sensing Accuracy
Physically-based retrieval converts radiance measurements into labeled training sets, resolving spatial de-correlation issues in land surface classification.
- Radar-Camera Fusion for Adaptive Obstacle Detection
A radar and camera fusion system uses preliminary scanning to identify obstacle types before high-precision tracking.
- End-to-End Vehicle Perception Model Training for Tracking Accuracy
An end-to-end trained perception model feeds detection outputs into tracking inputs, resolving occlusion-induced data switching errors.
- Autonomous Vehicle Emergency Vehicle Detection via Multi-Gate Light Analysis
Multi-gate light analysis identifies active emergency vehicles at distance, reducing false positives and collision risk.
- Object Recognition Using Feature Point Speed for Cluster Identity
A vehicle object recognition method maps sensor clusters to image regions and extracts feature points to calculate movement speed.
- Mobile Robot Control System for Dynamic Safety Limiting
A control system adjusts operation limitations on a mobile robot interface based on detected features of nearby individuals.
- Obstacle Detection Misalignment Adjustment
A misalignment detection unit assesses radar alignment with the vehicle longitudinal direction to adjust signal output conditions.
- Multi-Sensor Fusion Target Detection Reducing Computational Load
A multi-sensor fusion method projects millimeter wave radar regions of interest onto camera and lidar feature maps.
- Vehicle Travel Assistance Control for Abnormal Center Line Crossing
Camera-based travel assistance identifies preceding vehicles crossing center lines to generate braking and steering control signals.
- Driver Assistance Object Detection Performance Estimation
A driver assistance system estimates object detection performance using tracking and classification metrics to provide real-time reliability feedback.
- Human Detection Using Color Uniformity for Partial Occlusion
A human detection apparatus determines color uniformity in candidate areas to recognize obscured individuals.
- Traffic Light Signal Detection Using HSV Brightness Thresholding
A computer-implemented method divides traffic light images into signal portions and determines illumination status using HSV brightness values.
- ProtoFac Matrix Factorization for Deep Neural Network Interpretability
ProtoFac extracts meaningful prototypes from deep neural networks via matrix factorization, replacing internal layers with interpretable weighted concepts.
- Hybrid Crop Health Change Detection Using Vegetation Index and Structural Similarity
A hybrid crop health change image combines vegetation index differences with structural similarity measurements to quantify spectral and geometric variations.
- Mobile Vehicle Media Display Audience Estimation
A mobile vehicle media display system uses object sensors and neural networks to classify audience entities and compute engagement metrics.
- Vehicle Control Unit Environment Data Segmentation for Parallel Processing
Segmenting environmental data into independent records enables parallel processing within vehicle control units.
- Image Processing Device Partition Line Detection
A judging unit identifies improper regions using edge patterns to exclude them from target object detection.
- Dynamic Reference Range Adjustment for Driver Line-of-Sight Detection
A determination reference range changer adjusts detection parameters based on curve direction and relative vehicle position to prevent false alarms.
- Vehicle Collision Avoidance Path Display for Driver Alert Prioritization
An information display device generates and projects an avoidance path to guide vehicle navigation around detected obstacles.
- Autonomous Vehicle Obstacle Avoidance State Machine
A state machine determines target trajectories for autonomous vehicles by evaluating costs against thresholds to navigate around obstacles.
- Radar Locus Coincidence for Parallel Vehicle Recognition
A radar apparatus calculates the locus of representative points to determine object reality.
- Photoelectric conversion device, movable apparatus, photoelectric conversion method, and storage medium
A photoelectric conversion device uses a control unit to output signals between distinct accumulation periods within a full frame cycle.
- Sensor Data Preselection for Autonomous Vehicle Object Recognition
Transforming position information into a common coordinate system enables rapid object recognition while reducing bus system load.
- Decentralized Track Fusion Custodianship Model
A custodianship model distributes tracking functionality across multiple decentralized track fusers.
- Obstacle Recognition Device Sensor Bias Error Correction
A vehicle obstacle recognition device calculates an index value to identify objects across multiple sensors.
- Generative Transformer Model Retraining for Vehicle Object Detection
Retraining a pre-trained generative transformer model with unannotated vehicle sensor data to adapt the system for specific object detection tasks.