Back to datasets
Dataset assetOpen Source CommunityDriver Fatigue RecognitionEEG Data Analysis

SADT dataset, SEED-VIG dataset

The SADT and SEED‑VIG datasets are used for cross‑dataset driver fatigue recognition tasks via EEG data. These processed datasets can be used to test and train models to improve driver fatigue detection accuracy.

Source
github
Created
Jul 9, 2024
Updated
Jul 9, 2024
Signals
937 views
Availability
Linked source ready
Overview

Dataset description and usage context

EEG‑based Cross‑dataset Driver Drowsiness Recognition with an Entropy‑Optimization Network

Dataset Overview

The dataset is used for EEG‑based cross‑dataset driver drowsiness recognition tasks. It includes two publicly available datasets:

  1. SADT dataset:

  2. SEED‑VIG dataset:

Model Description

The proposed model, Entropy‑Optimization Network (EON), adopts a two‑step strategy to separate unlabeled target‑domain data. First, samples are pushed from the source domain toward the decision boundary, then a self‑training framework gradually separates them (entropy reduction), fully exploiting latent patterns in the unlabeled data.

Experimental Results

The method was tested on domain‑adaptation tasks, achieving 2‑class recognition accuracies of 89.2% and 77.6%, outperforming baseline methods.

Need downstream help?

Pair the dataset with AI analysis and content workflows.

Once the source passes your review, move straight into summarization, transformation, report drafting, or presentation generation with the JuheAI toolchain.

Explore AI studio