A WAD-YOLOv8-based method for classroom student behavior detection

Abstract This paper proposes an enhanced YOLOv8 model to address the challenges of complex classroom behavior detection. The model effectively overcomes the limitations of the original YOLO backbone, including the restricted receptive field and insufficient multi-scale feature learning due to fixed...

Celý popis

Podrobná bibliografie

Hlavní autoři:	Lisu Han, Xuejian Ma, Mengna Dai, Lu Bai
Médium:	Artigo
Jazyk:	Inglês
Vydáno:	Nature Portfolio 2025-03-01
Edice:	Scientific Reports
Témata:	Classroom behavior detection YOLOv8 Dysample
On-line přístup:	https://doi.org/10.1038/s41598-025-87661-w
Tagy:	Přidat tag Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!