VINocr API 2.0 VINocr API 2.0 Live Demo


Trained on our huge collection of images with Tensorflow 2.X and YOLO V4, our deep-learning VINocr API 2.0 provides a powerful and reliable OCR technology designed to help your clients quickly and effortlessly capture VINs printed on labels, receipts, invoices, documents, etc.

VIN numbers can be captured through barcode scanning, QR-code scanning, or even manually typed in. Quite often though, VIN numbers are printed on paper. How do you capture these printed VINs automatically? Our VINocr API is an unique technology designed to minize the headaches that can be associated with trying to capture printed VINs.

OCR Technology: The Core Technology Behind Our Deep-learning VINocr API

OCR stands for optical character recognition. Our VINocr VIN-capturing technology enables you to extract a VIN from an image and convert it into a editable VIN string. There are many OCR applications available out there but they need to be downloaded and installed locally, and can be very expensive. And most of them are even not specialized in VIN capturing as VINocr is.

Unlike some other OCR tools, you do not have to optimize your images before uploading them. Behind the scene, our system does automatic image preprocessing optimizaton before the image is fed to the OCR engine. The system reduces background noise and adjusts the resolution, automatically.

If the text font is too small or the background noise is too high (for example a background image) then our OCR process will fail like every other OCR tool. However, if you think your image is good enough to get usable results you can help us to improve this service and send your images to us. We are always looking for real world examples to improve this service.

Please note that the output VIN from this service might require a further sanity check, since there are a lot of complications that can arise during the recognition and translation process. Still, proofreading your VIN will be a lot faster and easier than having to type it all in, one character at a time.

VINocr API 2.0 Examples

Our deep-learning VINocr API works reasonably well on still pictures. It has never been tested on videos.