Scene text recognition reads text that appears in photos and videos of real-world surroundings, such as street signs, packaging or number plates. Unlike classic OCR for scanned documents, it has to cope with skewed perspectives, changing light, unusual fonts and cluttered backgrounds. A typical approach works in two stages: first locating text regions, then recognising the characters within them.
Optical character recognition (OCR) converts text contained in images, such as scans, photos or faxes, into machine-readable text. Only then can documents be searched, copied or processed automatically. Modern OCR uses neural networks and can also recognise tables, layouts and, in some cases, handwriting. Typical uses include invoice processing, archive digitisation and reading forms; quality depends heavily on image quality.
Computer vision is a field of AI in which computers interpret images and video. Typical tasks include detecting and locating objects, reading text in images and measuring components. The technology is used, for example, in quality control in manufacturing, in driver assistance systems, in logistics and in medical imaging.
Image recognition is a field of computer vision in which software identifies what an image shows – objects, people, text, logos or scenes. It is usually based on neural networks trained on large numbers of labelled images. Applications range from quality control in manufacturing and sorting product photos to supporting medical imaging. With images of people, data protection and potential bias require particular attention.
