[ad_1]
The descriptors will facilitate the recognition of objects and people.
Black, blue, brown, gray, green, orange, pink, purple, red, white and yellow. These are the eleven basic colors that artificial vision systems usually use to describe the tone of an image, clbadify objects or recognize and track people. Now, an international team of scientists, involving researchers from the Computer Vision Center (CVC) of the Autonomous University of Barcelona (UAB), has agreed on 28 names to describe as many colors so that the applications of the vision artificial can identify more tones and, therefore, better describe the different colors.
The new system, featured in Machine Vision and Applications magazine, incorporates turquoise, olive green, mint green, burgundy, lavender, magenta, salmon, cyan, beige, pink, dark green, olive, lilac, pale yellow colors , fuchsia, mustard, ocher, trullo, mauve, dark purple, lime green, light green, plum, light blue, peach, purple, tan and garnet.
"The names of the first eleven colors were chosen by linguists who studied different languages and published them in English, but linguists had no proposal to add more color names to artificial vision systems. hundreds that exist, so we looked for other ways to add these names and we designed this algorithm that does, "says Joost van de Weijer, CVC researcher.
The algorithm was formed with a base of 250 images extracted from Google for each of the 39 colors (11 + 28), applying statistical models to estimate the probability values, and learned to differentiate automatically and accurately each color.
Scientists have proved that the 39-color algorithm is not only useful for artificial intelligence, but its descriptors are also better valued by people. They did an experiment with volunteers who had to decide if the color and the name appeared on a computer screen, and their answers confirmed that to describe the colors, they preferred the system of 11 + 28 names instead of only 11. [19659009OtherfollowersLouYouighlightsthattheyhavedetecteddifferentdifferencesintheappreciationofthecolorsinrelationtotheobserver"Wehavenotpbadedtheoreticalandpsychologicalproblemsmaynothavedifferentiatedthanksthankseasily"hesaid
Source: La Vanguardia
