Imaging tests are not difficult.

As more and more people become experienced with "imaging studies," it is becoming a common belief that "imaging studies are difficult." Is this really true?
Probably few FlexInspector users would say that "image inspection is difficult" or "troublesome.
The biggest point is that "it is just doing what is natural. It's not a good product, so it's defective," that's all.
Many image inspections try to quantify and judge defects, and even recognize what defects are. To do so, they prepare complex algorithms and force users to make troublesome settings. In general, if we try to deal with defects, which we don't know where they occur, as a coping strategy, we will be stuck in a quagmire.
In the end, most image inspections are solved by "comparing with good products.
However, only when the variation of a "good" is so large that it cannot be compared, a recognition algorithm, such as taking dimensions, is necessary. In this case, however, the "good product" cannot be clearly defined, and the inspection itself will be out-of-date. In a sense, it can be considered "unsuitable for image inspection.