SBR-YOLO: Student Behavior Recognition Images for Smart Classrooms

A dataset of classroom images for student behavior recognition, referenced in a 2026 research paper. The dataset, named SCB-Dataset3, was used to evaluate the SBR-YOLO detection framework, which achieved a 74.2% mAP@50. The data was shared by author Yunming Zhang under a CC-BY-4.0 license.

Use Cases

• Training object detection models for student behavior recognition based on classroom images.
• Benchmarking attention mechanisms for handling occlusions and complex backgrounds as described.
• Evaluating feature fusion techniques for multi-scale object detection in educational settings.

Strengths

• Performance benchmark provided: SBR-YOLO achieved 74.2% mAP@50 on this dataset.
• Specific technical challenges addressed, including high intra-class variance and severe occlusions.

Limitations

• Dataset size is very small at 1.5 KB, suggesting it may be a sample or metadata only.
• Row count and column-level documentation are absent; data structure must be inspected after download.
• The description focuses on the model methodology; specifics on image count, resolution, or annotation types are unknown.

Provenance

Source

figshare, author Yunming Zhang.

Collection Method

Likely collected from classroom scenarios for research on behavior recognition.

Freshness

Last updated 2026-03-18 06:57:33.