| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide an object detecting device and method and a group learning device and method for quickening arithmetic processing at the time of learning and detection, and for detecting the target object in an arbitrary size, and for extremely improving discriminating capability at the time of detecting a target object by group learning. <P>SOLUTION: An object detecting device 1 comprises a scaling section 3 for generating scaled images by scaling down a gradation image to be inputted from an image output section 2, a scanning section 4 for sequentially manipulating and cutting out window images from the scaled images and a discriminator 5 for judging whether or not each window image is an object. The discriminator 5 includes a plurality of weak discriminators that are learnt in a group by boosting and an adder for making a weighted majority decision from the outputs of the weak discriminators. Each of the weak discriminators outputs an estimated value to estimate whether or not the window image is an object by using the difference of the luminance values of two pixels as a featured value. The discriminator 5 suspends the operation of computing an estimated value for a window image that is judged to be a non-object by using a preliminarily learnt suspension threshold. <P>COPYRIGHT: (C)2005,JPO&NCIPI |
