With Roxy AI, image generation AI is standard equipped. You can use it at no additional cost!

Learning data is essential for AI training. The accuracy of AI will definitely improve if we can collect more diverse learning data. However, it is difficult to collect many samples for products that rarely have defects.
Roxy AI's image generation AI can generate a variety of images as follows, which can reduce the challenges of sample collection:
- Generate defective images by specifying the shape and type of defect
- Specify the location where defects occur
- Modify the shape of existing defects
- Reuse an image generation AI trained on one product to generate defective images for another product (it can also generate defective images for products that have not yet started production and have no defects)
Of course, normal images can also be generated. By generating and training similar images in areas prone to over-detection, we can reduce false positives.
Training the image generation AI requires 10 to several dozen pieces of learning data. However, since we can learn from defects that occurred in different products together, the number of defective samples that need to be collected for one product is not that high.
We can generate images that are difficult to distinguish from real ones. Please take a look at the sample images.


Inquiry about this news
Contact Us OnlineMore Details & Registration
Details & Registration