Deep Learning for Appearance Defect Inspection in Laptops: A Model Comparison

Daily, thousands of laptops pass through repair centers of electronic device manufacturers. During visual inspections, technicians manually identify potential cosmetic or structural damage, documenting findings with photos or videos. This process safeguards the repair center against claims related to damage incurred in custody. However, the manual approach is time-consuming and prone to human error. To address this, we propose a computer vision model to automatically detect appearance defects in laptops. Experiments were conducted using five deep neural networks: Mask R-CNN, SSD, Swin Transformer, YOLOv5, and YOLOv10. Our best-performing model achieved a mean Average Precision (mAP) of 70.2%, showcasing the viability of such application.