Gender classification based on human radiation frequencies / Mohamad Hushnie Haron

• Main Author: Haron, Mohamad Hushnie
• Format: Thesis
• Language: English
• Published: 2015
• Subjects: Microwave communication systems; Microwaves. Including microwave circuits
• Online Access: https://ir.uitm.edu.my/id/eprint/15716/1/TM_MOHAMAD%20HUSHNIE%20HARON%20EE%2015_5.pdf; https://ir.uitm.edu.my/id/eprint/15716/
• Institution: Universiti Teknologi Mara

It is commonly known that living organism emits endogenous electromagnetic radiation. These electric and magnetic waves in electromagnetic radiation have certain characteristics such as frequency, wavelength and amplitude. Based on one of these characteristics namely frequency, a research focuses on classification of human gender was conducted. The goal of this research is to classify human gender using human radiation measured in frequency. In this research, the radiation emitted from nine points on human body was measured and analyzed using mean and min-max normalization. The hypothesis of this research which states that the frequencies of male radiation are not equal to the frequencies of female radiation was tested. Next, feature extraction on nine human radiation points was performed using Pearson's, Spearman's and Point Biserial correlation. From these correlations, two groups of points were selected and extracted. The extracted points together with nine points were used in gender classification and classifier validation. The results show that group using Point Biserial correlation achieved higher accuracy compared to group using Pearson’s and Spearman's correlation. Hence, human radiation frequencies can be used for gender classification. The outcome from this research can be used in many applications such as biometric recognition, visual surveillance, social networking and etc.