Analyzing Advertisement Concern Based on Age and Gender Features

• Main Authors: Ko-Hsin Cheng, 鄭可欣
• Other Authors: Yeuan-Kuen Lee
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/74786577584041524329

碩士 === 銘傳大學 === 資訊工程學系碩士班 === 98 === Data collection about the effect on the advertisement spends a lot of manpower and material resources in market survey. In order to reduce the cost, this study develops an automatic pedestrians classify system for advertisements standing in the roadside. The system also evaluates advertisement concern from the pedestrian’s walking speed. The system catches the pedestrian’s movement by 3 webcams and extracts color and space features from a pedestrian’s hair, dress, speed, height and other characteristics of while walking using image processing methods. By these 14 extracted features, the system used fuzzy rules to determine the membership of 8 classes of age and gender. The advertisement concern is also evaluated into three degrees. Finally, the efficiency and effectiveness of the proposed methods have been showed by various experiments in real cases.