FACE DETECTION METHOD AND DIGITAL CAMERA
First Claim
1. A face detection method comprising:
- classifying a plurality of images, generated by enlargement or reduction of time-wise successively captured images according to a plurality of specific multipliers, into levels by image size of the images;
sequentially reading out at one read-out unit at a time the images classified into each of the levels, using as the read-out unit a size that is smaller than the smallest image size from the image sizes of the classified images;
carrying out face detection processing to detect images in which it is determined that a face is present in portions of the image divided up into the read-out units, wherein when the face detection processing is repeatedly carried out to the time-wise successively captured images the number of candidate levels of the images used for the face detection processing for the next image is set based on the level(s) of image size of images in which it has been determined there is a face present so that the number of candidate levels is less than the total number of levels.
1 Assignment
0 Petitions
Accused Products
Abstract
A face detection method is provided, including: classifying into levels time-wise continuously captured images by increasing/reducing the total number of pixels; selecting sequentially and reading out image data for all of the levels using read-out units of the same size of pixels, or a smaller size of pixels, as those of the image with the smallest size of pixels; carrying out face detection processing by extracting candidate levels in which face image data is present, based on the read-out image data for each of the levels; and, when repeating the face detection processing, setting the number of candidate levels for face detection processing from the second time onward as less than the total number of levels. A digital camera incorporating the face detection method is also disclosed.
-
Citations
20 Claims
-
1. A face detection method comprising:
-
classifying a plurality of images, generated by enlargement or reduction of time-wise successively captured images according to a plurality of specific multipliers, into levels by image size of the images; sequentially reading out at one read-out unit at a time the images classified into each of the levels, using as the read-out unit a size that is smaller than the smallest image size from the image sizes of the classified images; carrying out face detection processing to detect images in which it is determined that a face is present in portions of the image divided up into the read-out units, wherein when the face detection processing is repeatedly carried out to the time-wise successively captured images the number of candidate levels of the images used for the face detection processing for the next image is set based on the level(s) of image size of images in which it has been determined there is a face present so that the number of candidate levels is less than the total number of levels. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
Specification