A key factor for OCR is character segmentation. MVTec combines freely definable regions with powerful blob analysis and comprehensive filtering techniques. This makes it possible to reliably extract characters even from complex backgrounds. A single character is read in less than 0.1 milliseconds.
Typical applications include:
- Serial numbers and codes on tires or metal parts
- Dot matrix fonts on complex backgrounds
- Handwritten digits or custom fonts
- Label and document recognition with syntactic and dictionary-based autocorrection



















