
Image recognition is one of the tasks with which services are best coped with elements of artificial intelligence. IBM launched in test mode a project that allows any user to test the capabilities of the Watson cognitive system regarding image recognition.
The service works with photos or pictures (you can upload your files, specify the address of the picture on the Web, or try to work with a photograph provided by the company). If the recognition was successful, the system will indicate a list of objects, presumably depicted in the picture or photo, and also show the probability of each assumption. If you wish, you can create your own category of objects (specify an existing category, or create a new one).
Typically, image recognition works in semi-automatic mode — certain objects are highlighted in the picture that the user must already identify. But the IBM service (
available by reference ) independently performs such work using semantic classifiers created by the system itself and introduced by developers in the learning process. A cognitive system can be recognized not only by objects, but also by events or environmental conditions using characteristics such as color, texture, shape, or edges of an image.
')
The service also uses the IBM Multimedia Analysis and Retrieval System (
IMARS )
platform . This allows developers to teach Watson, as well as create applications for their needs based on the capabilities of the cognitive system. For example, a retailer can create a tag, a classifier related to his clothes (trousers of a certain brand). And then, if necessary, the service can identify such clothes, if a photo with it appears on the network. That is, the seller (or manufacturer) of clothes can see how popular his product is among Web users.

By the way, the service is already working in test mode in the interests of MP Maritime (delivery of goods by sea). Using IBM resources, this company is able to track large cargo ships on satellite imagery. The capabilities of the service can be tested in the
Watson Developer Cloud environment on Bluemix.
Access to the
API is here. There is also
documentation and fork of the service on Github (
here and
here ).