Algorithm that predicts the movement of the eyes

The model adaptive visual prominence Care Whitening (AWS) is an algorithm developed by researchers at the Center for Research Information Singular Tecnoloxías (Citius) (Spain) paragraph gives address the problem of excess information in the field of robotics.
 
This system predicts the strengths of the images that attract the attention of viewer a, what could be Useful para Increase Effectiveness of ads.
 
The AWS model provides a tool that not only allows you to filter the noise para its application in the different areas of mobile robotics, but also serves as reference for the detection of contradictory largest in a particular image, even if in movement.
 
The visual representation of esta m result is called "saliency map" Attractions and Shows More para El Espectador, as confirmed by several experimental studies with humans.
 
The model considers the fact that a simple glance is enough that being human para para romper Essentials superfluous at the scene of a Do Around interpreting So the elements in the environment according their relevance. This Is A para constantly surprising capacity use, and whose numerous everyday applications include, example, UN Driving car.
 
Imagine that the driver of the UN car approaching a pedestrian crossing, where pedestrians UN group one is available across the street. Given the situation that detects a few meters ahead, the driver brakes gradually, but suddenly, UN rider crosses his path.
 
Suddenly, despite the passage of pedestrians remains on the viewing angle, the driver's attention is focused on the image area to be analyzed paragraph avoid abuse, Spend the rest of the background scene.
Only the strong points of the images that come to Our eyes are crucial para The process of human vision, since we have the ability paragraph interpret all the data about one of our Occur.
 
Thus, of our visual system you are responsible for filtering, a high velocity, the captured information, regardless of how it is irrelevant for the purpose of making feasible its subsequent interpretation in the brain.
 
This paragraph eliminate human capacity environment unnecessary or redundant information known as' visual attention. "A seemingly simple and transparent to the individual (in this case, the driver), which is, sin embargo, the m result of a set of Extremely complex evolutionary mechanisms and Optical Process neuronal level.
 
Teach One UN Parr Each robot play [Img # 30141] One of these steps Parr ensure that they learn a contradictory identify most relevant of their environment also requires numerous and intricate techniques (some of them inspired by human models), which once built paragraph solve one of the most common problems in the field of artificial intelligence: the presence of noise (or too much information).
 
The AWS model allows filter noise and detect the most relevant of contradictory images. The work has led to this algorithm, developed in the framework of industrial vision of the center, which was described in 2013 as "the best ever" In a study of the University of Southern California. (Source: SINC)