JP2005160122A5 - - Google Patents

Claims (16)

Detecting means for detecting a position of a specific subject from the captured image by comparing a captured image obtained from the image sensor with an identification pattern;
An imaging apparatus comprising: a focus control unit that performs automatic focus control based on the detected position of the specific subject. Detecting means for detecting a position of a specific subject from the captured image by comparing a captured image obtained from the image sensor with an identification pattern;
An imaging apparatus comprising: an exposure control unit that performs automatic exposure control based on the detected position of the specific subject. Detecting means for detecting a position of a specific subject from the captured image by comparing a captured image obtained from the image sensor with an identification pattern;
An image pickup apparatus comprising: white balance control means for performing automatic white balance correction based on the detected position of the specific subject.   4. The imaging apparatus according to claim 1, wherein the detection unit detects a position of a specific subject by comparing the shape of an edge detected from the captured image with an identification pattern. 5. .   5. The imaging apparatus according to claim 1, wherein the detection unit detects a position of a specific subject by comparing a luminance distribution of the captured image with an identification pattern. 6.   The imaging apparatus according to claim 1, wherein the detection unit forms the identification pattern from a plurality of person images.   The detection means includes a first detection means using a multi-layer feedforward neural network that detects the position of a specific subject by comparing the captured image with an identification pattern, and the detection result of the first detection means 7. The imaging apparatus according to claim 1, further comprising: a second detection unit configured to detect a position of the specific subject from the captured image when it is determined to be incorrect.   The said 1st and 2nd detection means has an evaluation means which evaluates regarding the said identification pattern about the said captured image, and detects the position of a specific subject based on the evaluation value, respectively. The imaging device described in 1.   The multilayer feedforward neural network has a nonlinear input / output characteristic having a saturation level in both large and small directions in the intermediate layer, and the second detection means is configured such that the output of the cell in the intermediate layer is the saturation level. 8. The image pickup apparatus according to claim 7, wherein if it is not, the detection result of the first detection means is positive, and if it is at the saturation level, the detection result of the first detection means is erroneous. . A detection procedure for detecting the position of a specific subject from the captured image by comparing the captured image obtained from the image sensor with an identification pattern;
A computer-readable storage medium storing a program for causing a computer to execute a focus control procedure for performing automatic focus control based on the detected position of the specific subject. A detection procedure for detecting the position of a specific subject from the captured image by comparing the captured image obtained from the image sensor with an identification pattern;
A computer-readable storage medium storing a program for causing a computer to execute an exposure control procedure for performing automatic exposure control based on the detected position of the specific subject. A detection procedure for detecting the position of a specific subject from the captured image by comparing the captured image obtained from the image sensor with an identification pattern;
A computer-readable storage medium storing a program for causing a computer to execute a white balance control procedure for performing automatic white balance correction based on the detected position of the specific subject.   13. The computer-readable medium according to claim 10, wherein the detection procedure detects a position of a specific subject by comparing the shape of an edge detected from the captured image with an identification pattern. 13. Possible storage medium.   The detection procedure includes a first detection procedure using a multilayer feedforward neural network that detects a position of a specific subject by comparing the captured image and an identification pattern, and a detection result of the first detection procedure includes: 14. The computer-readable computer program product according to claim 10, further comprising a second detection procedure for detecting a position of the specific subject from the captured image when it is determined to be incorrect. Storage medium.   The said 1st and 2nd detection procedure includes the evaluation procedure which evaluates the said identification pattern about the said captured image, and detects the position of a specific subject based on the evaluation value, respectively. The computer-readable storage medium described.   The multi-layer feedforward neural network has a nonlinear input / output characteristic having a saturation level in both large and small directions in the intermediate layer, and the second detection procedure is such that the output of the cell in the intermediate layer is the saturation level. 15. The computer-readable medium according to claim 14, wherein if it is not, the detection result of the first detection procedure is positive, and if it is at the saturation level, the detection result of the first detection procedure is erroneous. Possible storage medium.