Anyline Advancing AI Development with Major Funding from Austrian Research Promotion Agency

VIENNA, March 26, 2023 – Anyline, a global leader in AI-enabled mobile data capture and data insights, today announced that it received €2.64 million euros administered through the Austrian research promotion agency – Österreichische Forschungsförderungsgesellschaft (FFG) – as a public grant over the last three years to fund research and development of Closed Loop Training (CLT) for Artificial Intelligence (AI) models that speed the time to accurately capture and digitize analog information. 

Over the last three years, Anyline developed CLT, which is the foundation for a state-of-the-art AI-based software – leveraging computer vision, optical character recognition (OCR) and machine learning to more quickly and accurately capture visual information digitally. Think about it as a smart, continuously improving methodology to increase the accuracy of AI models. The result of CLT is that creating accurate data capture solutions using AI models and computer vision has been shortened from months to days. 

“Anyline’s mission is to make it easy to extract value from all of the static, analog information that surrounds us just by using the camera on your phone,” says Christian Pichler, CTO at Anyline. “The funding allows us to overcome technical challenges in realizing our vision, so when an external organization shares our vision and is willing to contribute towards that innovation, it’s extremely validating. With FFG’s support, we reinvented our tech stack, industrialized and automated the way we train our AI models and drastically reduced the time it takes to deploy new solutions.” 

Accuracy is the biggest challenge when using computer vision and AI models to capture analog letters, numbers and symbols –such as ID cards, utility meters or tire sidewall information. Improving accuracy from 40% to 50% is much easier than improving from 95% to 97%, which is often referred to as the ‘long tail’ or ‘diminishing returns’ problem. Typically, larger amounts of real-world data are needed to increase accuracy. CLT intelligently generates the quantity and quality of variants from smaller data sets to produce higher accuracy in much shorter timeframes than when compared to standard methods of AI model training, or machine learning. Over the last three years, Anyline developed this CLT factory to create custom AI-based computer vision solutions with release-quality results that can be built within days.