Visual Preference Research
machine learning is an idea that feeding data to the machine model and it builds it own logic based on the data, and thus is able to generate results that approximate the aimed result. In this visual preference research, the goal is to train a machine model that can recognize people's faces. The model is named ResNet50, which is a Deep Convolutional Network model that is used to perform image classification task. The dataset includes 5500 people's face images that have a score of 1 through 5 of how attractive they are. Through using python, the dataset is preprocessed and the machine model is feed with those images. After 30 epoch of training and 200 times of validation set, the ResNet50 is able to score human face that closely approximates the scores of images were tagged.