This paper presents the application of ANN in gender classification for children in Malaysia. The study involved kinematic data from 31 healthy children aged between 6 to 12 years old. The joint angles of hip, knee, ankle and pelvic were obtained using Vicon Nexus motion system at Human Motion and Gait Analysis Laboratory, UiTM Shah Alam. From 36 gait features, only 8 gait features that significantly differentiate between boys and girls. The 8 gait features data were then fed into the ANN models to classify the gender of children. An addition of synthetic data was used to improve the network. From performance of ANN gender classification models, the best model for this study is ANN-SCG model with 9 hidden neurons. The result shows that the performances of the ANN classification model for original gait features data were increased by 86.42% of accuracy when the synthetic data were added. The study showed that ANN application required a large number of sample size in order to produce good classification model.
Zakaria, N. K., Jailani, R., & Tahir, N. M. (2015). Application of ANN in Gait Features of Children for Gender Classification. In Procedia Computer Science (Vol. 76, pp. 235–242). Elsevier B.V. https://doi.org/10.1016/j.procs.2015.12.348