| 摘要 |
The present invention relates to a method for distinguishing the gender of a client or a user from an image obtained from a camera of a CCTV, a store, a cellular phone, etc. in real time. The method comprises the steps of: receiving a face image; generating a multiple region image composed of two or more images in the face image; generating an HOG image for each of the multiple region image; generating a characteristic vector based on the HOG image; and inputting the characteristic vector in the learned random forest, and distinguishing the gender based on the output value obtained from a plurality of trees included in the learned random forest. By providing excellent recognition performance for various input images including various angles, resolutions, lighting, poses, etc., and processing various input images in real time, the present invention is used for compiling statistics of clients who enter the store, or is applied to a system for providing an automatic service according to the gender of clients, thereby providing a retail solution. |
