Beauty Prediction

Introduction

Facial beauty has long been an area of focus for people, scientists, physicians, and artists, especially since the rise of digital media, the normalization of plastic surgery, and the wide-spread use of cosmetics. While beauty may be subjective, that doesn’t stop people from trying to assess and measure it!

Artificial intelligence and computer vision have approached this subject of interest, and novel ML models can now learn concepts of facial beauty that are programmed into it and produce estimations based on those criteria

Objective

In this lesson, we’ll develop a utility empowered by a pretrained, deep learning model ResNet that can score selfie images based on the face attributes. The model we’ll use has already been published as a Caffe model. Intrinsically, the OpenCV library supports loading and exploiting neural network models of type Caffe.