AIbase
EN
Home
Driver Drowsiness Detection
D

Driver Drowsiness Detection

Developed by chbh7051
Driver fatigue detection model based on ViT architecture, fine-tuned on the UTA RLDD dataset with an accuracy of 97.5%
Image Classification
Transformers
Open Source License:Apache-2.0#Driver Fatigue Detection#High-precision Classification#ViT Architecture
Downloads 131
Release Time : 4/3/2023

Model Overview

This model is fine-tuned from Google's ViT-base-patch16-224-in21k model, specifically designed for detecting driver fatigue states. By analyzing facial images, it can accurately identify whether the driver is in a fatigued state.

Model Features

High Accuracy
Achieves 97.5% accuracy on the evaluation set, demonstrating excellent performance
Based on ViT Architecture
Utilizes the advanced Vision Transformer architecture, effectively capturing global image features
Few-shot Fine-tuning
Fine-tuned from a pre-trained model, requiring relatively small amounts of training data for good results

Model Capabilities

Driver fatigue state recognition
Facial image classification
Real-time monitoring and analysis

Use Cases

Traffic Safety
In-vehicle Fatigue Warning System
Integrated into vehicle systems for real-time monitoring of driver status
Effectively reduces the risk of accidents caused by fatigued driving
Fleet Management
Monitors fatigue states of long-haul truck drivers
Improves fleet safety management
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
EnglishįŽ€äŊ“中文įšéĢ”中文ãĢãģんご
Š 2025AIbase