Synthetic Driver Image Learning for Drowsy Driving Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional simulation methods for detecting careless driving have high accident risks and low accuracy due to limited simulation scenarios, making it difficult to verify driver states in real-world conditions effectively.
Innovation Solution
A simulation learning-based drowsy driving simulation platform system that generates synthetic images of drivers with varied features and environmental conditions, using deep learning to enhance the accuracy of careless driving detection and control vehicle functions accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional simulation methods are used to estimate driver gaze area, then the system can extract ground truth from simulation environment, but the accuracy of verification is low due to limitation of number of simulations
Solution Approach 1:
The patent creates synthetic driver images through image synthesis technology, copying and transforming real driver images to simulate various drowsy states. This allows generating unlimited simulation data without additional physical experiments, thereby improving detection accuracy while maintaining productivity.
Solution Approach 2:
The system pre-processes driver images to create a comprehensive dataset of drowsy driving scenarios before actual detection. By preparing synthetic images with various hair types, genders, ages, skin colors and expressions in advance, the system ensures high detection accuracy when actual driving occurs.
2Measurement precision
If verification is performed by creating driver careless state in actual vehicle driving situation, then real-world accuracy can be tested, but there is high risk of accident
Solution Approach 1:
The patent introduces synthetic driver images as an intermediary between real driving and actual driver state verification. These synthesized images serve as a safe mediator that allows verification of careless driving detection without exposing the actual driver to dangerous situations, thus eliminating accident risk while maintaining verification accuracy.
Solution Approach 2:
Instead of manipulating real drivers in actual vehicles, the system creates copies of driver images through synthesis and applies various transformations to simulate drowsy states. This copying approach enables safe verification of detection algorithms without any physical risk to drivers.
3Measurement precision
If the system applies image synthesis with multiple driver features (hair, gender, age, skin color, expression), then the accuracy of careless driving determination is improved, but the device complexity increases
Solution Approach 1:
The image synthesis module is designed to handle multiple driver features (hair, gender, age, skin color, expressions) through a unified transformation framework. This multi-functional approach allows the same system to generate diverse synthetic images without requiring separate processing pipelines for each feature, thus improving detection accuracy while controlling complexity.
4Reliability
If the smart cruise control processing part controls steering and braking in addition to general functions, then vehicle safety is improved, but the device complexity increases
Solution Approach 1:
The smart cruise control system dynamically adjusts its control functions based on the detected driver state. When drowsiness is detected, the system automatically activates additional steering and braking controls. This dynamic adaptation allows the system to maintain high safety standards only when needed, avoiding unnecessary complexity during normal driving conditions.
Data Source
AI summary
Disclosed is a simulation learning-based drowsy driving simulation platform system for detecting careless driving in conjunction with deep learning, the simulation learning-based drowsy driving simulation platform system comprising: a drive state warning device configured to determine a driver's careless driving from a captured image, determine a driver's careless driving determination level, and output the determined level; a smart cruise control interworking part configured to transmit the driver's careless driving determination level outputted from the drive state warning device; and a smart cruise control processing part configured to control a vehicle according to the driver's careless driving determination level transmitted by the smart cruise control interworking part, during a smart cruise control operation.