How to deal with true angle and predict angle has 180° error?

[Nioolek]

When I predict a text area, rotate image and crop it, I often find that if a text area`s angle is 225°, the rrpn predict 45°. After I input the crop image croped by 45°, the crnn result goes wrong. So , can author or ohers can advise some methods to make sure whether the correct prediction is 45° or 225°.

[BaofengZan]

@Nioolek Hi, have you run through the code of rrpn? I have some questions : 1. Where is the code of skew-nms? 2. When training MSTRA-TD500, there is no test result output at the end? Have you encountered this kind of problem?

[Nioolek]

@BaofengZan Hello, At present, I only use the network provided by the author to make predictions, and plan to train RRPN network suitable for Chinese in the next two weeks. I think you should be Chinese. You can add my WeChat : asda1231 ,for further discussion.