The photos from a website containing photos with self reported height and weight information.

http://www.cockeyed.com/photos/bodies/heightweight.html

 

BMI 예측 관련 프로젝트에서.

Introduction

This project implements the idea presented in the following paper:

2013, Wen, Lingyun and Guo, Guodong

A computational approach to body mass index prediction from face images

The paper propsed that a person's body mass index (BMI) can be estimated from facial features without measuring the actual height and weight. A trained feature detector based on the active shape model (ASM) is used to extract points of interest on the face; then several facial features are calculated using the points. After features are computed on the dataset, a model is trained using machine learning approaches.

Data

The photos used in this project are crawled from a website containing photos with self reported height and weight information.

(http://www.cockeyed.com/photos/bodies/heightweight.html)