Fig. 1

Flowchart of the DAL-YOLOv8 framework. In the first stage, the Cantonese embroidery images are augmented and annotated. In the second stage, the model detects objects and locates their positions. In the third stage, the output results are analyzed.