Images_Creation_Using_Videos

By Stan He, 11th Nov. 2019

Purpose: Create a series of images for deep learning using videos recorded on cellphone

Note: If you used iPhone to record the objects to be learned by CNN, then you need to convert from MOV to MP4 first before extracting images from the original videos.

• Step One: Download critical dependencies for video manipulation If you are using Mac, please type the following command in the terminal:

• brew install ffmpeg

• Step Two: After the successful installation, run the following command for video batch conversion from MOV to MP4 if you are using iPhone to make the recording.

• Step Three: If the recorded video is in MP4 format, then you are good to go.

• 3.1 To make the Shell script executable, in terminal, RUN:

• chmod +x video2images.sh

• 3.2 If you would like to create 5 images every second of video (ie every 200ms), then in terminal, RUN:

• ./video2images.sh /home/user/directoryOfVideos 5