Synth.Eye GAN — Industrial Inspection Dataset

Synth.Eye GAN — Industrial Inspection Dataset Training dataset for the Synth.Eye GAN project — synthetic YOLO-format images of industrial parts with and without fingerprint residue defects, generated by three StyleGAN2-ADA models and a physically motivated compositing pipeline. The dataset drives two downstream YOLO models: one for part orientation classification and one for fingerprint defect detection. Both models are available at LukasMoravansky/Synth-Eye-GAN. Dataset… See the full description on the dataset page:

Source: Hugging Face Hub (LukasMoravansky/Synth-Eye-GAN-Data). Metadata imported from the dataset’s Hub tags.