我上传了一些png动图，但是DSP报错，报错的内容为
错误：不支持 / 的操作数类型：‘int’ 和 ‘NoneType’
想请教一下如何解决这个问题
Hi @xbersis,
Could you tell me for which of your projects you encountered this issue so I can have a look please?
Regards,
Louis
你好，
如何告诉你是哪个项目，就告诉你项目名称吗？
项目名称是：动作识别
谢谢
Hello @xbersis,
Got it thanks.
I just noticed that you have only one label in your dataset.
To be able to use the image classification, you need more than on label otherwise the classification results will always be the same even with images that are completely different.
Could you try to add another class such as unknown?
Regards,
Louis
好的谢谢你，
当时在提问的时候，我选择的处理块是“光谱分析”，这导致了报错。
我想请问一下如果是想视觉上学习一个动作，比如用摄像头检测挥手，我是应该选择哪一个处理块呢，需要上传什么类型的数据呢？
At the moment we do not support videos or animated images.
@janjongboom has been working on something similar recently but a lot of testing / implementation still have to be done so I don’t know about the ETA.
An interesting thought would be to substract the difference between two images to only get the resulted “action” (considering you are using a fix camera) so it would remove part of the background and then create a new generated image.
Note that the Transfer Learning Image block won’t work in that case because it has been pre-trained on “real” images. You’ll also need to implement your own NN architecture.
Regards,
Louis
非常感谢您的回复，给我带来了很大的帮助 
2 Likes
我一定要尽快学中文 