| Summary: | 碩士 === 國立暨南國際大學 === 電機工程學系 === 102 === In recent years, biological gender identification is a hot topic. It is widely applied to the entrance security. As face detection technology becomes mature, how to use it to accurately determine the gender, is a popular research in the field of computer vision.
In this paper, we propose some technologies that can reduce the amount of data generated by the Moment which is used as gender identification characteristic feature extraction tool. These technologies can enhance the recognition results.
This system contains pre-processing module, feature extraction module and classification identification module. The images are input to this system. The pre-processing module detects a face with OpenCV. The detection result will be segmented into different sizes for the internal, general and external features. Through various Moment processes such as Geometric moments, Zernike moments, Eigenmoments, and ICA Moments, features can be extracted. Then, we use SVM for classification.
This paper uses the FERET database for the experiments. It shows that the best classification results can reach 86% by ICA Moments with SVM classification.
|
|---|
