Amazon announced during the event AWS re: Invent three new services Artificial Intelligence that allow developers to build apps that can recognize images, convert text to voice and understand natural language. Rekognition, Polly and Lex these are their names, they were designed by the experts of the US company using advanced deep learning algorithms. The services are available only in a few countries, but these are to be global distribution.
Rekognition is the service that analyzes the images and recognizes faces, objects and scenes. A human being is able to easily understand the difference between a dog and a cat, the computer can perform the same task only after “trained” the neural networks . This requires time and computing power. Amazon has cataloged billion images, then simply use the Rekognition API to write apps that identify not only the animal kind, but also its breed. The technology of course also works with people, of which are identified sex and facial expression. The service can be exploited in various ways, for example from the surveillance cameras.
Polly is the service that allows you to create “talking” applications, converting the text to speech ( text-to-speech ). Polly offers 47 different voices (male and female) in 24 languages. The developer can implement the real-time streaming audio or save the converted text in MP3 or Ogg in high quality (22 kHz). Data is encrypted and audio is transmitted via an SSL connection.
Lex finally is the service that allows a computer, smartphone or tablet to listen and respond to the user. The technology is the same used to Alexa, the personal assistant of the speaker Echo and tablet Fire and then uses advanced algorithms of speech recognition and natural language understanding . With Lex you can also create chatbot for Facebok Messenger.