More than words

Juan José Moreira

Speech recognition is a discipline that amis at allowing a device to recognize and understand spoken words. Until recently, the idea of holding a conversation with a computer seemed reserved exclusively to the realm of science fiction movies , however , things are changing. Since that first system from Bell Labs named Audrey, developed over 60 years ago and which was only able to recognize digits, voice recognition technology has evolved and today the number of people who “talk” to their cell phones to dial a number, send an SMS, set an alarm or perform a search on the Internet is increasing .
But speech recognition is also opening to other fields of application. The new TV models of Samsung allow controlling many of their functions through voice. From classic functions such as increasing or decreasing the volume or changing the channel , to the most advanced ones , such as asking the TV to recommend us a program or to list all movies of a specific genre being broadcast at that moment.
In the automotive world, many brands are also starting to incorporate this technology into their vehicles. For example, Ford with “Sync” and Opel with “Intellilink” . These systems are used to control through the voice certain car systems such as the air conditioning, the hands-free to make calls or the GPS navigator.
Speech recognition is also expanding into the field of call-centers management tools where it is used to guide the client through the various options when making a call, for monitoring the service provided by the agents or even to measure the degree of customer satisfaction.
These are just a few examples of how voice recognition systems are showing as an emerging technology with a growing ability to penetrate different market sectors. But voice can be used to extract more than words and get to know, for example, the mood of the speaker. Thus, there are several research projects aiming at the developing of a virtual agent able to interact naturally with a human through a combination of speech recognition and emotion recognition .

In GRADIANT we support these technologies and we are currently engaged in developing our own solutions for voice recognition and recognition of emotions in voice within the framework of different projects.