Identification systems objects, faces and landscapes are in an early stage of development, but soon will be more effective than people, according to an executive of Facebook.
The robotic facial recognition, and advanced technology, is on track to surpass the human eye precision, once 3D images in the processing of the data, a challenge in which technologists work, coming one told Efe the creators of the identification system that uses Facebook.
Eden Shochat, co-founder, acquired three years ago by Facebook to improve facial recognition massively, recently visited Madrid to participate in the Future Trends Forum (FTF).
Accuracy levels of computers thanks to artificial intelligence in identifying faces are almost equal to those of the human eye, but could soon surpass those of men, says this young Israeli, currently a partner of an investment fund endowed 164 million for venture capital projects.
Shochat said that the main obstacle to the identification of robotic images do not already have an accuracy of one hundred percent, above the man, whose effectiveness is around 98 percent, not technical. The specialist noted that the algorithms would be solved, and the main shortcoming is the limited availability of 3D images.
The facial recognition technology works from comparisons between the attributes of the face of the person to be identified and millions of individuals with similar characteristics contained in a huge database.
The greater the volume of information available, the greater the accuracy of the system when it comes to recognizing faces, details.
Since the 3D cameras are still too expensive, three-dimensional images are now accessible to the few, which is expected to change as time passes, said Shochat, who has created several successful start-ups before selling.
The use of 3D images of faces massive facial identification to be a technological milestone, added the specialist. Their applications could be extended to areas such as security, for the automatic recognition of criminal or criminals from comparing images.
The key breakthrough in the last years of this kind of inspired by the use of artificial intelligence is due to the momentum of the “deep learning” and the ability of computers to teach themselves technologies, following models of neural networks.
The entrepreneur points out that it is producing a real “tectonic shift” in this area, with capacities by the increasingly close to those of human machines.
In fact, remember, the machines and create productions with hints of creativity, such as pieces of music with a quality comparable to the works of the best artists.
In the coming years, thanks to artificial intelligence, the advertising industry will be able to set “a la carte” messages when directing your prospects and know more precisely and in advance what they need or want.
In fact, even better machines and interpret radiographs that some physicians in diagnosing diseases such as cancer, and computer systems are able to detect errors in industrial processes that allow anticipate potential manufacturing problems and, thus, to optimize production costs.