Gender recognition technology of whole body image

There are numerous potential applications for the Internet of Things (IOT) at the present stage. In the topic of image processing, the gender recognition usually adopts face information to identify the gender of a person. Therefore, if the input image loses face information, it will result in the wrong identification result. In this paper, we proposed a multiple-attributes (MA) recognition method. Not only facial parts, but also use gender of wearing characteristic for gender identification. The paper also develop an automatic segmentation algorithm to locate the wearing characteristic. The objects features will be effectively extracted, and classified into some object attributes by classifier. Finally, the attributes are given to the corresponding weights to obtain male scores and female scores. According the gender scores to determine the gender of people. This paper has multiple-attributes (MA) to solve the defects of the face attribute recognition.