The best mobile text recognition SDK

Would you recommend this product?
No reviews yet
Hey everyone! We are super happy to be featured on Product Hunt today! Big thanks to @__tosh for the hunt! :) My name is Daniel and I’m the CTO at Anyline! Anyline is a mobile OCR (Optical Character Recognition) SDK, which can be implemented in any kind of mobile application. We have been working hard for the past years to enable developers the recognition of a variety of text, codes and numbers for their mobile app use cases! We’d love to hear your opinion, questions and feedback on Anyline! :) >> Today and the whole rest of the week, we are giving away a 10% voucher on the Anyline Do It Yourself Basic Edition which is perfect for Startups & SMEs, looking for the perfect, easy & customizable OCR solution! (voucher code: ScanAllTheThings!) << Looking forward to hear your feedback! :)
Upvote (12)Share
So how does it work? I simply snap a picture with the mobile camera and Anyline processes the image?
Hi Kevin! Thanks a lot for your question! Actually we took a different approach with our Anyline SDK. The scanning happens in the video stream and without snapping a picture. Through processing video frames the SDK tries to extract the text more or less in real-time. Therefore the result can be more accurate as the SDK compares results from single frames (aka pictures) with each other and gives back the result with a higher confidence. All of this happens within milliseconds as the processing stays directly on the device.
Can I use Anyline's scanning functionality within a Chatbot?
Hi @3x14159265, I thought you would come up with a bot question ;) At the moment Anyline does not work within Chatbots. Our offline focus keeps us from the possibility to integrate it, but in the future we're definitely planning on making this happen!
What happens when the internet connection is bad? Does this influence the results or quality of the scans?
Hey @reiter_jakob ! Great question, thank you! :) Anyline does not need a server connection in order to process the images. It works completely offline on the device! This means you receive instant feedback while scanning and a close-to-real-time result! :)
@sabrinarueck Offline processing is a great approach. How does it perform on old/weak devices, especially on Android?
Hey @stefanraffeiner ! Of course the SDK works better on new or high-end devices but we support Android back to Android 4.0.3 ! On old devices with a slow hardware it could take longer for the result, but if the device's camera works properly, you will definitely get a scan result!
What kind of technologies are at the base of the SDK? (Deep Learning etc.)
Hi Rod! Great question - as we've put a lot of time into finding the right approach, this is definitely one of my favorite questions! We started with well-known Computer Vision and Text Recognition algorithms: openCV & Tesseract which we enhanced with our own pre-processing technology. By now we also made a free Open-Source Tool for developers who want to use Tesseract (it got also hunted a few months ago: Throughout the last year though, we put a lot of effort into using Deep Learning on mobile devices. The challenge definitely was to make Deep Learning frameworks run efficient and fast without adding lots of MB to the apps. Currently we're using TensorFlow and a see great results in accuracy.