Project Description

IdeaNova Technologies is a software company with over a decade of experience in IT security and secure video streaming. They specialize in developing software for Inflight Entertainment Systems, healthcare, and other industries. For this project, IdeaNova wanted a program that could detect scene transitions and end credits within movies so that users could easily scrub through the content of a video by selecting individual scenes as well as skip to the next episode of a series when the credits appear. This would be implemented into IdeaNova's existing inflight entertainment software which allows flight goers to watch movies and TV shows during their flight.

Our Solution

For our solution, we decided to use a variation of Amazon's ShotCoL algorithm which focuses on finding the similarity between different shots and using that information to group different sets of shots into scenes.  For reference, a shot is a series of frames captured by a single camera uninterrupted, and a scene is a series of shots that make up a cohesive part of a story. For the work of splitting a video into shots, we used the library pyscenedetect, which given a video input, outputs a series of images, each one representing a separate shot. Then, we used OpenCV to compare shots to each other and find scene boundaries. More info on Amazon's ShotCoL algorithm can be found here.
 

Backend Technologies:

• Python
• Pyscenedetect
• OpenCV

Frontend Technologies

• React