摘要:
PROBLEM TO BE SOLVED: To surely recognize an object better than before.SOLUTION: First recognition processing recognizes, for example, a hand. Second recognition processing recognizes, for example, a person. When an obtained image is inappropriate to recognize a hand because, for example, the hand in the image is too small, the second processing is performed to recognize a person in the image. When the person is recognized, an image quality parameter such as a zoom rate is newly acquired. Then, an image having the changed image quality parameter is acquired and the first recognition processing is performed again. By repeating the process until recognition of the hand, the hand is surely recognized. The invention can be applied to, for example, a personal computer.
摘要:
PROBLEM TO BE SOLVED: To provide an apparatus performing image recognition wherein processing speed is improved. SOLUTION: A pyramid image generation part 20 generates, from an inputted image, a plurality of images having different resolution (scale coefficients). Filter processing is applied to the images having the different scale coefficients. A filter is a filter of a Gaussian function of Gaussian width σ=1, for example. Though the filter processing is applied to the plurality of images, processing can be reduced because not performing the processing by use of a plurality of filters having different Gaussian width. This invention can be applied to a recognition apparatus recognizing an object from an image or a learning apparatus performing learning for the recognition apparatus. COPYRIGHT: (C)2010,JPO&INPIT
摘要:
A storage portion stores, in association with first pixel values, table values including either gamma-corrected values obtained by performing gamma correction on the first pixel values using a predetermined reflectance component or second pixel values calculated based both on the gamma-corrected values and on the first pixel values. A pixel-value generating portion includes an extracting portion and a determining portion. The extracting portion extracts at least one of the table values corresponding to a pixel value of a subject pixel. The determining portion determines a pixel value of an output image based on the at least one of the table values extracted by the extracting portion. The predetermined reflectance component is the reflectance component calculated by the reflectance-component calculating portion when the pixel value of the subject pixel is substantially identical with the average luminance of the peripheral pixels.