Video indexing and search technology takes a leap with kPoint


Today, video accounts for more than 60 percent of traffic on the Internet. But despite the high number of videos on the World Wide Web, we are often faced with situations where we cannot find the right video, and more importantly, get to the right point within them. So, at a time when everything on the internet – including web pages, documents and email, is searchable, why should videos be any different?

In-video search becomes all the more significant in the case of enterprise videos where the average length of a video could vary between eight to ten minutes. ComScore’s survey conducted in January 2014, validates this fact, finding that the average viewer engagement drops by 60-75% after 4 minutes and 20 seconds.


Browse through this video to experience kPoint’s powerful search capabilities.
Use the search widget on the top right corner and type any key word from
the video to get to the exact point in the video where it appears

Traditional video platforms search only for titles, tags, and comments within videos, thereby limiting the effectiveness of search. In a maze of videos with similar sounding terms, it becomes a challenging task to locate the precise bit of knowledge that you are looking for.

But the days of poking and prodding within videos are now over. kPoint scans every frame of every video so well that users can search through every spoken word or important text on screen to find precise points where a phrase was said or shown. kPoint automatically indexes all videos in your library and assists the author in generating highlights at key points in the video to enable a truly rapid and efficient browsing experience.

By fine tuning the indexing of video content and making it available for search, kPoint makes in-video search a reality. kPoint auto generates its intelligence about content in a video from 3 main sources – 

  • Spoken word analysis using Automatic Speech Recognition (ASR) technology to recognise each word spoken within a video.
  • Computer Recognition using Object Detection for understanding the content of an image and Optical Character Recognition (OCR) technology to recognise text within images in the videos.
  • Slide content ingestion technology to index text within slides or documents used in the videos.

Apart from a super-efficient video viewing experience, this technology also finds application in better categorization of content and intelligent suggestions based on a viewer’s preferences.With fine grained indexing, we have made video browsable like a webpage.