python+opencv人脸识别训练报错及解决方法

1.AttributeError: module 'tensorflow' has no attribute 'variable_scope'

解决：

将原代码：

import tensorflow as tf

改为：

import tensorflow.compat.v1 as tf

若未能解决，重新下载tensorflow

.whl文件：Links for tensorflow

源代码：https://github.com/tensorflow/tensorflow

注：若下载源代码文件，将原来的tensorflow文件替换为下载好的tensorflow文件即可

2.AttributeError: module 'cv2' has no attribute 'face'

解决：需要opencv-contrib-python，在win+R中输入cmd进入windows命令行程序，输入pip install opencv-contrib-python，并检查opencv-python和opencv-contrib-python是否同版本

3.Empty training data was given. You'll need more than one sample to learn a model. in function 'cv::face::LBPH::train'

解决：原来是我的项目中，cv2的 imwrite仅支持png和jpeg格式，将jpg改成png即可