This project introduces a diffusion-based generative framework guided by bounding boxes to synthesize high-quality industrial images along with corresponding segmentation maps. The method is designed ...