Deep learning for construction waste detection using ConvNeXt V2 EMA attention and WIoU v3 loss

Puntos clave

• Enhanced accuracy in waste detection was achieved using ConvNeXt V2 attention mechanism.
• The model reached an accuracy of 92.3% on the test data, highlighting its effectiveness.
• Evaluation utilized deep learning techniques with WIoU v3 loss to improve performance metrics.
• This approach may enable more efficient waste management practices, benefiting construction processes.