download fer2013dataset(第一步：下载fer2013数据集)
- Use pandas to parse the csv file, and then save the original image data as a jpg file, classify it according to the purpose and label, and save it to the corresponding folder.(第二步使用pandas解析csv文件，再将原始图片数据保存为jpg文件，并根据用途和标签标签进行分类，分别保存到对应文件夹下)
- run fer2013process.py PrivteTest folder/ PublicTest folder/Training folder
- Build model.py(新建model.py)
- Build utils.py(新建utils.py)
- train.py
- Download the trained model haarcascade,simple_CNN(下载好训练模型haarcascade,simple_CNN)
- Build utils.py (编写utils.py)
- Build image.py and input a picture to recognize all the faces and expressions in the picture (编写image.py输入一张图片可以识别图片中所有人脸及表情)
- real-time detect(编写video_test.py)