I'm sorry...

We cannot guarantee the performance of FlexInspector's inspection capabilities.
Users have commented that the FlexInspector has a much better inspection capability than our previous products.
One reason for this is that we believe that FlexInspector is the ultimate visual inspection method and that it is difficult to do better.
We statistically process product variations, set the sensitivity at the optimum value that balances detection capability and yield, and use all image information to make a judgment on the extracted defect candidates. This results in a result that is quite close to that of visual inspection.
Although there is still room for improvement in operability and external interfaces, we believe that the algorithm is now in a state of completion as an algorithm for "visual inspection" using image processing.
Another point is that the inspection capability of a visual inspection system is determined by the "user's sense. Even if the same detection capability is demonstrated, it will be evaluated differently if the person in charge changes. The manufacturing department will focus on yield, and the quality control department will focus on quality. Of course, the delivery destination may also give a different evaluation. It is difficult to "guarantee the performance" of something of this nature. That is why we sell only after prior performance evaluation by the users themselves.
So...
Inevitably, our sales destinations are limited to end-users who are able to evaluate the products themselves.
It is very difficult for trading companies that only have a "slip of paper" to handle these products.
We also want to avoid inquiries led by equipment manufacturers without evaluation by end users. Even if a large inspection system is made, we cannot guarantee the quality of the product.
We will make a product that the end-user can give us a go-ahead based on a preliminary evaluation. That is all we can think about.