add feature: images can now be loaded from the bytes() object

[SerinaNya]

• add feature: images can now be loaded from the bytes() object
• update README_cn.md
• have NOT updated README.md yet

[guofei9987]

Thanks for this. A merge request is always welcome.
But I do not get it why bytes is needed. It was designed to read 3D-matrix image. See this https://github.com/guofei9987/blind_watermark/blob/master/examples/example_no_writing.py

[SerinaNya]

If only 3D-matrix images can be passed in, the user must convert the image to something like a cv2 object first. If the wrong parameters are used when reading the image using cv2, the blind_watermark will not work properly.

In my case, your lib is used with a web framework which returns the image uploaded by the user as a bytes object. Therefore I created this PR.

如果只接受 3D-matrix 对象，那么使用者必须事先将图片转换为 cv2 对象之类的。如果使用 cv2 读取图像时提供的参数错误，将导致你的库无法正常工作。
我是将你的库和 Web 框架一起使用，Web 框架会将用户上传的图片以 bytes 对象的形式返回给我。