History of Technology 16
Natural Language Processing
Natural language processing allows the car to read and understand the languages used by humans. Many researchers hope that a sufficiently powerful system language processing could learn on their own, by accessing databases on the Internet. Some applications are derived from this information indexing services and computerized translation.
Robotics is close A.I. because it is necessary that robots can navigate and manipulate objects and solve sub-problems locating adjacent (to know where it is), mapping (to learn what's around) and movements and route plotting (to know how to get there).
Artificial Perception is the ability to use the product input sensors (cameras, microphones, sonar, etc.) to deduce various aspects of the world. Computer Vision is the ability to analyze visual input. Special problems are found in sub-problems such as speech recognition, facial and objects.
Social Intelligence means that the agent is able to predict the actions of others, by understanding the motivations and emotional states. This involves elements of game theory and decision, as well as the ability to model human emotions and the perceptual skills necessary to detect them.
For a decent human-machine interaction, it must present her emotions, or at least appear polite and sensitive to the people who interact with it.
Software for visual recognition
Visual recognition software can analyze an image of the human face to estimate age, based on the signs of aging, US researchers say.
The software was developed by a team at the University of Illinois.
"Measuring the years is very difficult," says Dr Thomas Huang. "If you use the face to estimate age can learn apparent age, or how old a person looks."
Researchers trained by using software algorithms 1,600 different people, each with 5 images, resulting in a total of 8000 images. Their ages were between one and 93 years.
Looking girls and uses computer software to determine which features are best determined from a person at a certain age.
An image of size 100 x 100 pixels 000 pixels results 10, all having a color level. Comparing how dark or how light it each pixel to other pixels, the software estimates the apparent age of the individual.
"A woman who is a makeup will look younger," said Huang. "A fine texture of the face will register a younger age."
Software records and face shape. The relative position of the eyes, nose, ears, mouth shape, all change over time and indicates a person's age.
"If we estimate age person with a margin of error of 10 years, then the accuracy is about 80%," says Huang.
Shape, position, color and texture clues support not only older. But also race, sex, even emotions.
The software can recognize and positive facial expressions such as smiling
This is just an academic exercise in computer science. Face recognition software will be useful for fast food companies who want to know how many young people buy a sandwich or some type of clothing stores that will help them bring in their store various types of clothing specific to age and sex.