Ashwanikumarkashyap / context-based-video-seek-google-cloud-intelligent-video

The project gives the ability to seek videos to a particular point/portion based on the key context provided by the user. Based on a user query (object, genre, category, etc.) the project filters out the videos that have the key context in them and also highlights the portion in the video that where it appeared. The project utilizes Google's Cloud Video Intelligence, Cloud Speech-to-Text API, Google Cloud Storage via Google Cloud Platform, and using Node.js for the integration of everything.

Motivation

• I wanna see a part of a Friends episode amongs all season where Brad Pitt shows up.
• I wanna see the airport battle in Captain America: Civil War.
• I wanna only watch the heated moments in a football match, nobody likes the fillers right?
• Or anything random like, I wanna watch penguins from a video about Antarctica on NatGeo!

But the issue is, that nobody remembers the actual timestamps. You don't know when your desired content occurs in the video. Have you ever thought it would have been better if we were able to jump to our favorite portion of the video without actually seeking manually in the video? Wouldn't it be cool if you just type/say what you want to see and automatically get the location to that part of the video?

Solution

The solution is here! Presenting the most innovative, out of the box solution where you just type/say what you want to see and this project filters all the videos having the context of your search query. Our project filers the all the videos and highlights the portion of the video where the searched context appears in that video. No more manual seeking to the timestamp is required.

The project utilizes Google's Cloud Video Intelligence, Cloud Speech-to-Text API, Google Cloud Storage via Google Cloud Platform and uses Node.js for the integration of everything.

This solution will not only help build a better customer experience for video streaming sites like YouTube, Netflix, Amazon Prime Video etc., but would also help in better categorization of the content, that can be recommended to the users in future. It also helps filter explicit content with the provided kids mode. Not to forget the portion of the topic you want to study a day before the exam from the MIT OpenCourseWare.

Future Scope

An important usecase one can think of moving this project forward is in the surveillance sector. It has endless possibilities in this area. We will be able to get information, tag them, store them and use them whenever we want. It will reduce putting in a lot of efforts in manually going through the videos.

It's the one stop solution to all the problems. So what are y'all waiting for.