Clipping

• After we have tagged each second with relevant topics and subtopics, the next step is to clip the video into segments based on the topics.
  Read more in Docs > Architecture > Machine Learning.
  
  

Algorithm

• We use a sliding window to detect changes in the topic of a video, Comparing with a sliding mean or median to compare the topics of the current window to the previous window.
• If the difference between the two is greater than a threshold, we consider the topic to have changed.
• We then save the start and end times of the window right before the topic changes.

We then use these start and end times to split the videos, save raw files to Amazon S3, and save metadata to our API.