Digital camera having subject judgment function
원문보기
IPC분류정보
국가/구분
United States(US) Patent
등록
국제특허분류(IPC7판)
H04N-005/228
H04N-007/18
G06K-009/00
출원번호
UP-0973617
(2004-10-27)
등록번호
US-7605847
(2009-11-10)
우선권정보
JP-2003-366319(2003-10-27)
발명자
/ 주소
Chiba, Toru
출원인 / 주소
Hoya Corporation
대리인 / 주소
Greenblum & Bernstein, P.L.C.
인용정보
피인용 횟수 :
0인용 특허 :
7
초록▼
A subject judgment program is capable of reducing the calculation load on a computer functioning as a device for judging the subject in a subject image. In a controller of a digital still camera, when a CPU running the subject judgment program (stored in a flash memory) obtains image data from an A/
A subject judgment program is capable of reducing the calculation load on a computer functioning as a device for judging the subject in a subject image. In a controller of a digital still camera, when a CPU running the subject judgment program (stored in a flash memory) obtains image data from an A/D converter via a first interface circuit, the CPU generates two-dimensional distribution data of product-moment correlation coefficients for each of model image data stored in the flash memory by successively calculating the product-moment correlation coefficient between the model image data and each part of the obtained image data, identifies a piece of model image data corresponding to two-dimensional distribution data having the highest maximum value, and thereby identifies subject information which has been associated with the identified model image data.
대표청구항▼
What is claimed is: 1. A computer-accessible recording medium storing a program comprising computer-readable instructions that cause a computer to function as: a storage module which stores a plurality of pieces of reference image data in a storage device associated with subject information definin
What is claimed is: 1. A computer-accessible recording medium storing a program comprising computer-readable instructions that cause a computer to function as: a storage module which stores a plurality of pieces of reference image data in a storage device associated with subject information defining subjects in images represented by the respective image data of the plurality of pieces of reference image data, an input module to which subject image data is input, a distribution data generation module which generates pieces of two-dimensional distribution data of product-moment correlation coefficients for each of the plurality of pieces of reference image data stored in the storage device by calculating the product-moment correlation coefficients between each of the plurality of pieces of reference image data and parts of the subject image data, an identification module which identifies a piece of two-dimensional distribution data having the highest maximum value among all the two-dimensional distribution data generated for all the pieces of the reference image data, and an output module which reads out the subject information associated with the reference image data corresponding to the identified pieces of two-dimensional distribution data from the storage device and which outputs the subject information, wherein: the distribution data generation module includes: a color space conversion module which converts color space of the subject image data into YCrCb while converting color space of all the reference image data stored in the storage device into YCrCb, a coefficient calculation module which calculates product-moment correlation coefficients between Cr components of each reference image data and parts of Cr components of the subject image data and thereby generates two-dimensional distribution data of the product-moment correlation coefficients regarding the Cr components for each reference image data, and which calculates product-moment correlation coefficients between Cb components of each reference image data and parts of Cb components of the subject image data and thereby generates two-dimensional distribution data of the product-moment correlation coefficients regarding the Cb components for each reference image data, and an average calculation module which executes an average calculation process, for calculating each geometric average of the product-moment correlation coefficients regarding the Cr and Cb components at the same two-dimensional coordinates, for each reference image data and thereby generates two-dimensional distribution data of the geometric averages for each reference image data, and wherein the identification module identifies a piece of two-dimensional distribution data of geometric averages having the highest maximum value among all the two-dimensional distribution data of geometric averages generated for all the reference image data. 2. A digital camera, comprising: a storage unit which stores one or more pieces of reference image data associating each reference image data with subject information which defines a subject in an image formed by the pieces of reference image data, an image pickup device which picks up a subject image formed by an objective optical system and thereby successively generates subject image data, an image processing unit which changes image quality of an image displayed on a display device according to the subject image data by executing computation to the subject image data using proper correction values, a distribution data generation unit which generates pieces of two-dimensional distribution data of product-moment correlation coefficients for each of the one or more pieces of reference image data stored in the storage unit by calculating the product-moment correlation coefficients between each piece of reference image data and parts of the subject image data, an identification unit which identifies a piece of two-dimensional distribution data having the highest maximum value among all the two-dimensional distribution data generated for all the pieces of reference image data, a readout unit which reads out the subject information associated with the reference image data corresponding to the identified two-dimensional distribution data from the storage unit, and a setting alteration unit which alters the correction values used by the image processing unit into correction values corresponding to the subject information read out by the readout unit, wherein: the distribution data generation unit includes: a color space conversion module which converts color space of the subject image data into YCrCb while converting color space of all the reference image data stored in the storage unit into YCrCb, a coefficient calculation module which calculates product-moment correlation coefficients between Cr components of each reference image data and parts of Cr components of the subject image data and thereby generates two-dimensional distribution data of the product-moment correlation coefficients regarding the Cr components for each reference image data, and which calculates product-moment correlation coefficients between Cb components of each reference image data and parts of Cb components of the subject image data and thereby generates two-dimensional distribution data of the product-moment correlation coefficients regarding the Cb components for each reference image data, and an average calculation module which executes an average calculation process, for calculating each geometric average of the product-moment correlation coefficients regarding the Cr and Cb components at the same two-dimensional coordinates, for each reference image data and thereby generates two-dimensional distribution data of the geometric averages for each reference image data, and wherein the identification unit identifies a piece of two-dimensional distribution data of geometric averages having the highest maximum value among all the two-dimensional distribution data of geometric averages generated for all the reference image data. 3. A method of identifying a type of a subject image, comprising: preparing a plurality of pieces of reference image data respectively representing a plurality of different types of images, a number of pixels of subject image data being different from a number of pixels of each of the pieces of reference image data; scanning, of a piece of reference image data and subject image data, the piece of reference image data and the subject image data having a lower number of pixels, within the other of the piece of reference image data and the subject image data to evaluate similarity therebetween at every predetermined scanning position to obtain two dimensional distribution data from product-moment correlation coefficients between each of the pieces of reference image data and the subject image data at every predetermined scanning position, the scanning being repeated for each of the plurality of pieces of reference image data to obtain the two dimensional distribution data for each of the plurality of pieces of the reference image data; identifying one of the plurality of pieces of the reference image data most similar to the subject image in accordance with the plurality of pieces of the two dimensional distribution data corresponding to the plurality of pieces of the reference image data; converting color space of the subject image data into YCrCb while converting color space of all the reference image data into YCrCb; calculating product-moment correlation coefficients between Cr components of each reference image data and parts of Cr components of the subject image data and thereby generating two-dimensional distribution data of the product-moment correlation coefficients regarding the Cr components for each reference image data, and calculating product-moment correlation coefficients between Cb components of each reference image data and parts of Cb components of the subject image data and thereby generating two-dimensional distribution data of the product-moment correlation coefficients regarding the Cb components for each reference image data, and executing an average calculation process, for calculating each geometric average of the product-moment correlation coefficients regarding the Cr and Cb components at the same two-dimensional coordinates, for each reference image data and thereby generating two-dimensional distribution data of the geometric averages for each reference image data, wherein the identifying comprises identifying a piece of two-dimensional distribution data of geometric averages having the highest maximum value among all the two-dimensional distribution data of geometric averages generated for all the reference image data. 4. The method according to claim 3, wherein the number of the pixels of each piece of the reference image data is smaller than the number of the pixels of the subject image data. 5. The method according to claim 3, wherein the numbers of the pixels of the plurality of pieces of the reference image data are the same. 6. The method according to claim 3, wherein the preparing includes generating each of the plurality of pieces of the reference image data by reducing the number of pixels of a plurality of pieces of predetermined model image data. 7. The method according to claim 3, wherein the preparing includes reducing the number of pixels of an object image to generate the subject image. 8. The method according to claim 3, wherein the identifying identifies one of the plurality of pieces of the reference image data which corresponds to the two-dimensional distribution data having the highest maximum value of any two-dimensional distribution data as the reference image data corresponding to the subject image.
연구과제 타임라인
LOADING...
LOADING...
LOADING...
LOADING...
LOADING...
이 특허에 인용된 특허 (7)
Ray, Lawrence A.; Nicponski, Henry, Face detecting camera and method.
Ii,David L.; Reitz, II,Elliott D.; Tillotson,Dennis A., Method and computer program product for identifying and correcting systematic noise in a pattern recognition system.
※ AI-Helper는 부적절한 답변을 할 수 있습니다.