Using Crowdsourced Metadata to Improve Search in Video

By Patrik - July 8, 2015 (Last updated: April 3, 2017)

With the increasing amount of video being produced and published everyday, new tools to search and find parts of any video will be needed. Today I stumbled over one of those new tools.

Say that you are looking for some video material you want to watch, then realising that the video is longer than you expected, and you only want to see parts of it (how many presentations have you endured that never get to the point?). Today pretty much the only way you can do that is by skipping back and forth in a desperate attempt to locate what you are looking for.

ClipMine has a really cool solution to this, using the crowd as well as automatic algorithms to tag the video, and then creating a table of contents that can be seen in the player. The metadata will make the video players content-aware and allow better findability for video. You can find an introduction to ClipMine in a post on their blog.

At Vidispine we love all things video and metadata, and are looking forward to see how this unfolds. I see a huge potential for building applications on top of the ClipMine metadata, especially if they open up an API for third-party developers.

What would you do if you had deeper access to the content in a video?