Human Attribute Prediction using Deep Face

Face Attribute Prediction including Age, Gender is an interesting topic among many researchers and is one of the most interesting activities in pattern recognition, personal computer interaction. Most applications use this process because a person’s face is considered a very rich source of information. This paper not only predict Age and Gender but also predicts Emotions and Racism from a featured image using a deep face method. Dataset used for training the system is FER (face emotion recognition) from kaggle. Fer2013 contains approximately 30,000 facial RGB images of different expressions with size restricted to 48×48 and the main labels of it can be divided into 7 types of classes namely angry, disgusted, fearful, happy, neutral, sad, and surprised. For testing IMDB and Wikipedia data set are used. IMDB WIKI data set is the largest publicly available facial data with gender, age, and name. IMDB contains 460,723 images and Wikipedia contains 62,328 images. The overall accuracy achieved by the model is 97.23 percentage which was considerably high as compared to the previous models