Inspiration
A lot of students rely on videos online for lectures. These lecture videos are often long and it is hard to tell if a specific subtopic is present in the video. This inspired us to build an interface to facilitate deeper search within videos.
What it does
This provides an interface to search for the presence of a string within a video and also to point out its time-based locations.
How we built it
We used Google Cloud Speech-to-Text API to obtain audio transcripts of a given video URL. The user inputs the video URL to be searched for and a search string. The transcript is parsed to find the string and its timestamp to output whether the string is found in the video, and if so, a GUI which points to which parts of the video. Python was used predominantly for dealing with the Google Cloud Service and Javascript was used for web interfacing.
Challenges I ran into
First time hackers, no previous experience with "front end" development.
Accomplishments that I'm proud of
Implementing an interesting idea in a short time span and working on 4 hours of sleep!
What I learned
Javascript, HTML, interfacing them with Python and using Google Cloud Services.
What's next for InfinightVideo
Implement this for a set of videos instead of searching through just one. Possible application would be to search for an episode by just remembering a dialogue.
Log in or sign up for Devpost to join the conversation.