Image Recognition is a technique for identifying objects, buildings, people and patterns in photographs. In image recognition, the images are first processed using image processing and then the objects to be recognized are extracted.
Image recognition technology has been used for many years in industrial applications, for example in the recognition of parts and components in industrial production. Apps for smartphones and tablets have also opened up image recognition to consumer technology. This allows users to capture logos, trademarks, cars, people or albums and have them analyzed in a cloud or using Google's Visual Search.
The applications of image recognition are extremely diverse, ranging from biometric feature recognition such as face recognition, hand recognition, iris recognition, and retina recognition for personal identification; object recognition for augmented reality and location-based services; 2D code recognition and analysis; mark recognition for industrial production, merchandise control, and printing; and traffic sign recognition in automobiles. Since objects can look completely different from different perspectives and under different lighting, object recognition often turns out to be extremely difficult and is supported by artificial intelligence( AI), machine learning and deep learning. Another problem is the separation of object, lighting effects and background. While human perception solves these problems unconsciously, powerful computers need programs to separate between object and background.
Image recognition methods always work with image databases, which they search for comparable patterns.