Saving Face
Research has shown that it takes one-tenth of a second for a person to look at your picture and form an impression about your personality.
A recent study investigated several low level features such as saturation, hue and contrast as well as high level features such as face size, setting, dress style, glasses, smile, background, head tilt etc. that influence our impression of images.
Saving Face looks at some of these elements and provides users with feedback about what they can do to improve the quality of pictures.
We first used Microsoft’s Cognitive Service Vision API, in particular the Face API that detects human faces and returns face attributes containing machine learning-based predictions of facial features. Of the available features, we used low level features such as exposure, noise and blur and high level features such as emotion (anger, contempt, disgust, fear, happiness, neutral, sadness and surprise), smile, and accessories such as sunglasses. Additionally, we used Google’s Logo Detection API that detects product logos within an image.
Log in or sign up for Devpost to join the conversation.