We usually do two things with a video. We watch it or we create. The latest Microsoft Garage project promises to change all that. For those who don’t know, Microsoft Garage is the skunkworks lab for experimental projects that breathe life into fantastic ideas. Video Breakdown went live last month.
Video Breakdown is a new platform from Microsoft that makes any video you upload searchable. On the new video platform, you can search for videos on a specific topic and jump to the best segments of a video with a click. Microsoft calls it Cognitive Video Indexing.
Here’s how it works.
The web tool analyzes the uploaded video file and indexes its contents using a range of powerful Microsoft services. The index contains information like an audio transcript, face tracking with identification, differentiation of speakers, optical character recognition, and even sentiments projected in the content. So, the powerful backend services “break down” a video to its core ingredients and give you many ways to access the information in it.
Now, you can search the video using the web app.
The interface gives you many filters to dive into the depths of a video.
You can search a 3-hour video by a single word spoken by a presenter. Go straight to the segment where a speaker mentions a specific topic. Looking for a term used in a text slide within the video? Video Breakdown uses OCR to help you jump to that part. Try searching for a celebrity by name in a video you upload. The analysis even understands the level of positive vs. negative sentiments expressed in the spoken or written content.
You can upload a video and create your own breakdown for others to use. Use the controls in the online editor to interact with the video.
An Innovative Tool You Can Use Right Now
The Microsoft Blog pinpoints the benefits if it is adopted on a bigger scale.
This could be a useful tool if you’re hosting a conference, or any event with speakers. It’s also a big jump from how you search now, where video content is sorted by channels or metadata – and very subjective manual tagging.