V! Studios develops high volume solution for extracting text from scanned documents faster and cheaper for NASA’s Johnson Space Center by utilizing Amazon Web Services (AWS) Lambda serverless technology.
V! Studios, a creative technologies company, has developed a new high-speed, low-cost cloud-based solution for optical character recognition (OCR) and text extraction. Standard OCR software extracts text as a page-by-page serial task while the V! Studios solution coordinates an ephemeral server to process each page. Using hundreds of ephemeral servers operating in parallel, the system completes a scan for hundreds of pages in minutes.
V! Studios developed its text extraction engine “SOCRATES” (Serverless Optical Character Recognition And Text Extraction System) to assist businesses and enterprises to quickly identify and discover information within document repositories. The product provides a quick and cost-effective service, enabling text extraction in a cloud workflow while maintaining data security. Upon learning of V! Studios’ efforts, NASA realized the potential of the technology to solve an existing problem.
The Extravehicular Activity (EVA) group at NASA’s Johnson Space Center faced the challenge challenge of converting a massive amount of spacesuit safety and test documentation into an easily searchable database. Although their existing OCR procedure was cloud-based, the serial text extraction process significantly slowed the workflow.
“When I learned V! Studios was developing a serverless OCR solution, I immediately saw the potential benefit for a current problem in our cloud-based data integration application. By decoupling the OCR text extraction function using Lambda’s serverless framework within GovCloud, we can now extract text from documents in a fraction of the time and cost.” – Cuong Nguyen, NASA EVA Project Manager
SOCRATES is an automatic, serverless compute system that scales instantly to process a thousand page document in minutes while avoiding many of the requirements of a conventional cloud computing environment. With V! Studios’ solution, NASA’s text extraction tasks gets the exact amount of compute power at exactly the time needed, eliminating the cost for idle time.
V! Studios CTO, Chris Shenton, enthuses, “The cool thing about SOCRATES is that it’s event-driven and it instantly handles all the scale-out for you; you pay only for active use, so the cost-savings are huge – it’s a game-changer.”
V! Studios looks to deploy SOCRATES as a SaaS (Software as a Service), with availability coming to the AWS Marketplace in the first quarter of 2018 in all U.S. regions including GovCloud. The API will provide quick and cheap text extraction service for any organization to integrate into their cloud workflow.