Discover Your Island University

Graduate Projects


Project ID: 429
Author: Shuyi Zhao
Project Title: An Android Application for Facial Aesthetic Analysis
Semester: 3 2014
Committe Chair: Dr. Scott A. King
Committee Member 1: Dr. Dulal Kar
Project Description: An Android application FaceAnalysis is designed to analyze a human face from a photograph. Users can take a photo by a custom camera or choose a picture by the Google Gallery. Then the application analyzes the input image and give some improvement advice about the input face to users. To achieve this function, three problems need to be solved: face detection, facial feature point detection and the aesthetic analysis. Both face detection and facial feature point detection are achieved using OpenCV. The face detection uses a Haar Cascade classifier and a LBP Cascade classifier. In the application, an image is checked by LBP Casecade classifier first. It is faster and costs less system resources, which is important for mobile devices. Then if there is no face detected by LBP Cascaded classifier, Haar is used due to its high accuracy. The facial feature point detection uses ASM algorithm. Two aspects of Aesthetics is considered: the symmetry of a face and the Golden Ratios on a face. The UI of the FaceAnalysis is designed as a multi-pane layout, which adapts for different size screens. The FaceAnalysis has been tested on animal photos, beautiful human photos, ordinary human photos, non-ideal human photos, and the group with multiple faces.
Project URL:   429.pdf