CG generation for defect classification (teacher images for machine learning)
We will generate defective images such as scratches, bruises, and cracks that can be used as training data.
In the field of manufacturing, high-speed automation of visual inspection using image sensors is advancing. On the other hand, the images that serve as training data for defect extraction require many different patterns, and optimal settings for cameras, lenses, and lighting tailored to the target are also essential, leading to numerous challenges. In response to customer inquiries about whether "3DCG can solve this," our company offers a solution for generating defective images for defect judgment (training images for machine learning) that creates various defective images such as scratches and missing parts. The types and application methods of defects, such as scratches, dents, and cracks, can be freely set using parameters. Additionally, the settings for the light source can be customized in terms of quantity, color, irradiation range, and angle, making it possible to closely resemble real environments. We anticipate responding to the detection of various abnormalities in a wide range of inspections. If you have any challenges regarding diverse images of scratches and defects, please feel free to contact us.
- Company：シリコンスタジオ
- Price：Other
