Long Range Video Image Of Gender Identification

As an accessory means of identity recognition, gender identification technology is playing an increasingly important role in anti-terrorism, national public safety, etc. Especially in intelligent video surveillance system for long range identity recognition, gender recognition technology of long range is of greater research value. Since data in the video surveillance system are mainly video images, this study conducts a research on gender identification in light of long range video images.Firstly,on the basis of extensive relevant literature gender recognition technology at home and abroad, and through characteristics analysis of long range video images,the study carries out a research on gender recognition technology of long range,by making use of information on H-GEI and on clothing texture variations; Secondly,the study extracted and processed video data from data set B of CASIA, and indoor gender identification database of long range was established; Thirdly, by making use of gait energe image, combined with Fourier descriptors, HGF features were figured out, which showed that HGF features can be taken as those used to identify genders in long range video images, and several significant conclusions were reached; At last, by making use of local binary pattens, combined with texture block similarity matching, we take clothing texture variation features, TSV. Experimental results showed that TSV can be regarded as gender recognition characteristics in long range video images, which embodies robustness to be excellent for gender recognition under multi-view variations.