Defective Image Generative AI technology developed by RUTILEA, Inc. - solves training data shortage for visual inspection automation
RUTILEA announces the development of an AI-based defective product image generation technology.
A bottleneck in the automation of visual inspection using AI is the lack of defective product images to serve as training data. To solve this problem, RUTILEA has developed a technology to generate a large number of good/defective images from a small number of good/defective images using a generative model based on deep learning. The diffusion model is characterised by its ability to generate high-resolution images.
This enables a large number of defective images to be generated from around 10 defective images, which can then be used to improve the accuracy of visual inspection systems.
RUTILEA is in the process of incorporating this defective image generation AI into its 'No-code AI image processing software' ImagePro, which is expected to be available by June.
(Reference) No-code AI image processing software ImagePro