Age prediction

There are multiple ways to predict age from a given input photo. Earlier methods work by calculating the ratios between different measurements of facial attributes such as eyes, nose, mouth, and so on. Once facial attributes are calculated based on their size and distance, ratios will be calculated, and age categorization will be done using rule-based engines. Here comes the problem: this method may not work perfectly when we don't have the perfect full frontal face photo that we typically see on a profile picture on any social media platform.

There are multiple ways to predict and identify facial features. One such method is Gaussian Mixture Models (GMM), which were used to represent the distribution of facial patches. We then moved to super vectors and Hidden Markov Models (HMM) to represent the facial patch distributions. The best performances were showcased by employing Local Binary Pattern (LBP) and the dropout Support Vector Machine (SVM) classifier.