We got into a discussion about YouTube videos and how it would be a good idea to have a Ctrl-F capability for audio.
What it does
Through an Web application interface, users can type the word and the video they want to search. It finds occurrences of the word in an XML file of a YouTube video transcript and returns the times it occurs. The video is then viewable on the webpage and the user can skip the video to the time they want.
How we built it
Challenges we ran into
We spent a lot of time trying to figure out a way to preprocessing the audio file into text due to lack of existing API that can help our speech recognition search.
Accomplishments that we're proud of
Fortunately, we found an alternative way by using video transcripts with selenium; and finally Yoogle is working (YES!!!).
What we learned
We came from different backgrounds in tech. Some of us never work on any hackathon before, so, we had a chance to learned more about Git, HTML and CSS, while others who have more experience in hackathon, learned more on how to use YouTube transcript and speech audio processing.
What's next for Yoogle
Apply an automatic speech recognition and Machine learning to be able to make the search more robust and able apply for any video (Not only Youtube video). We want to create an API for Yoogle, so that we can provide a plugin for MOOCs or other online courses.
Log in or sign up for Devpost to join the conversation.