Human Face Detection Following by Gender and Age Estimation Using Patch Based Discrete Cosine Transformation and Histogram Oriented Gradient Features
Automatic human age estimation has become a challenge in the field of Machine Learning and Computer Vision. In this research, both the age estimation and gender classification work flow can be divided into three modules such as Face Pre-Processing, Feature Extraction and Classification. In the Face Pre-Processing stage, to detect face from a given input image, Viola and Jones Face Detector was used. The main challenge comes when the feature extraction is needed. This feature extraction needs to be precise so that it can show the differences between other classes. So far, for this step, the detected face was separated into patches and then discrete cosine transformation and histogram oriented gradients were applied so that the system compares the same feature while training. While creating model, if the redundant features are considered then it effects the system to predict precisely. At last, for the classification section, SVM was applied to get an acceptable accuracy. The proposed approach was applied on CACD database as well as IMDB database with the age range of 16 to 60.