YOLOv8 Dataset: Annotated Passenger Images for Double-Decker Bus Monitoring

Binwei CHEN's dataset supports the manuscript 'Strategic Approaches to Real-Time Passenger Monitoring in Double-Decker Buses Using YOLO Technologies'. It contains 412 raw and annotated training images, an independent test set of 80 images, and machine-readable CSV files for all tables and figures. The data enables full reproducibility of the reported computer vision experiments.

Use Cases

• Training object detection models based on the 412 annotated images with rectangular bounding boxes.
• Benchmarking model performance on an independent test set based on the External_Dataset_80.
• Conducting learning curve experiments based on the SetA_0.1.csv to SetA_1.csv files.
• Reproducing research results based on the provided Table and Figure source data in CSV format.

Strengths

• Includes a temporally distinct external test set of 80 images collected from 2025 06 16 to 2025 07 14.
• Provides full reproducibility data with source CSV files for all 10 tables and figures from the manuscript.
• Released under a permissive CC BY 4.0 license, facilitating reuse.

Limitations

• Column names and detailed metadata for the image files are not provided in the input.
• The dataset scale is moderate, with 412 training and 80 test images.
• The input provides conflicting information on the 'Rows' field (unknown vs. None).

Provenance

Source

Binwei CHEN, associated with the manuscript 'Strategic Approaches to Real-Time Passenger Monitoring in Double-Decker Buses Using YOLO Technologies'.

Collection Method

Likely collected via camera systems on double-decker buses for passenger monitoring.

Time Range

Test data collected from 2025 06 16 to 2025 07 14.

Freshness

Last updated 2026 03 24.

Geography

null