scene classification
Modification on https://github.com/GKalliatakis/Keras-VGG16-places365
Download model from the urls, and ready to go.
- python vgg16_places_365.py
Result is a txt file, top 5 classification of each picture.
Use classify.py can move pics to its category folder.
Other files:
test.py cut the video based on critical frame.
pyscene.py Extract all frames from a video.
places365 Folder contains files to do scene classification.
classify.py Move pics to its folder based on classification result.
场景分类
在https://github.com/GKalliatakis/Keras-VGG16-places365 基础上
进行了修改和整理了环境配置
下载模型,然后就可以跑了
- python vgg16_places_365.py
结果是txt文件,每个图片的前五类别。
用classify.py 可以把图片按文件夹分类。
QAQ 找了好久场景分类的与训练模型
其它文件:
test.py 是根据关键帧密度剪切视频
pyscene.py 是提取视频中所有的场景图片
places365 是进行场景分类的文件夹
classify.py 是根据识别的结果将图片移动到对应的文件夹中
全部配置好的文件:
链接:https://pan.baidu.com/s/1iXLX104YwT249fyMLVSFnw
提取码:xllz