Image recognition is a technology that is used to analyze images to detect text, people, locations, buildings, product inventory, production quality assurance, medical anomalies, and more. Image recognition is commonly used with machine learning, a type of artificial intelligence. Image recognition is used in ecommerce for product recommendations and product search.

Consumer applications for image recognition include mobile shopping apps that help shoppers find products online by scanning items in-person from a physical store. This method of image driven price matching is commonly referred to as showrooming. Image recognition like this is found on consumer mobile devices and is available through the device camera.

Another common example of image recognition is Optical Character Recognition (OCR). OCR software scans a typed document, image, or handwritten correspondence to detect and output recognized text. Advanced OCR methods use neural network algorithms to distinguish between light and dark patterns to present a best guess of the scanned image.

Image recognition is commonly used with a convolutional neural network (CNN), a deep learning architecture which does not require manually entered test data. A CNN is used for pattern matching, object recognition, and facial recognition. A convolutional neural network does more than image recognition work, it is also capable of classifying audio and signal data. A CNN can also be programmed for computer vision tasks such as object recognition as is used in real world autonomous robots and vehicles applications.

Many businesses can benefit from image recognition capabilities, for example:

  • Sort and analyze large numbers of images quickly. Image recognition, when used for machine learning, can perform high volume sorting and analysis much faster than manual methods.
  • Quality Assurance Automation. Improved quality assurance processes can be made through automated image recognition.
  • Efficient Item Classification. Warehousing operations can use image recognition technologies to sort and classify products quickly.
  • Extend Working Capacity on the fly. Insurance adjustment agencies can use image recognition technologies to speed claims processing to human adjusters. This efficiency is especially helpful after emergency situations where large numbers of the population are affected.