In the fall of 2015 Josh Kanner, CEO of SmartVid, had an interesting problem in a space not traditionally computerized: people on construction sites need to document what work is being done each day for legal and compliance reasons as well as to show progress to executives in charge. Construction sites are hazardous to navigate, making it a chore to even begin. But Josh sensed that software could help. Given the huge growth in video-capture and inexpensive storage, could Rocket help build a robust front-end for their new platform?
Rocket jumped in and worked alongside the Smartvid team to build a powerful video processing application that intelligently processes massive amounts of video content. We helped integrate a hierarchical tagging system that allows users to quickly organize many videos and communicate with others in real-time. We integrated an ASR (Automatic Speech Recognition) system that automatically tags videos with the words spoken in them which drastically reduces processing time. And we delivered on time. Our work helped construction companies reshape the tasks on the job site, making them far more efficient and safe.
Building simple video annotation and sharing tools into Smartvid helped project managers communicate efficiently. We helped this both manual and automated ways:
• Users can draw on any frame in a video to point out where something is wrong. They can also add comments for even more detail, allowing in-depth communication where there was none before.
• We implemented ASR (Automatic Speech Recognition) and then auto-tagged certain words throughout a project. This led to massive speed gains in team communication.
The way an app is organized can mean the difference between success and extreme frustration. With the Smartvid app we sweated the details of workflow to help users find and manage their videos easily.
We built in multiple ways to find videos, each satisfying a different use case. We exposed tags and categories at every level so that people could organize things how they wanted. We also auto-organized videos based on tags and generated thumbnails of these giant videos so users could quickly navigate around libraries with ease.
In the difficult environment of low bandwidth and large videos, we made the SmartVid app faster in several ways:
• Users can capture and upload multiple videos at once. This greatly reduces the time needed to get videos into the system.
• The application uploads videos once, and only once, into their final video slot. Following this simple principle slows down many video platforms which end up moving files later on, adding unecessary and debilitating latency.