Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Facial Relative Ranking

No description
by

Fiona Yang

on 11 December 2013

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Facial Relative Ranking

Facial Relative Ranking
Introduction
Detect certain face features from input face image
Calculate their similarities to pre-evaluated faces in our face database
Decide the attractiveness of the input face.
Library & Environment
Flandmark: Used to abstract 8 distinct points for different parts of a face
Algorithm Breakdown
Ranking
Nose size
Mouth size
Face height/width ratio
Eye size/Face size ratio
Eye size/Mouth size ratio
Eye size/Nose size ratio
=>
Weighted facial similarity
Testing Result
Yawei Wang
Kaiyu Yan
Ningyuan Yang

Matlab implementation of Face Detection, Pose Estimation and Landmark Localization in the Wild (FDPELL): Used to extract 68 general points of a face
Training
Construct a faces database by dividing variety of faces into 10 ranking categories (1 - 10)
For each face image, extract the parameters for the eight feature points using Flandmark library as the following picture shows:
Use Matlab FDPELL code to extract 68 general points for each face image.
Based on the eight feature points, use adaptively learning to group the 68 general points to different facial parts and generate the raw data for each face.
Calculate all needed benchmark parameters using the raw data acquired from previous steps and stored into our database by category.
Testing
Extract the feature points and calculate all benchmark parameters using Flanmark libraries and Matlab FDPELL Code.
Compare customer’s parameters with data stored in each of the categories
Determine the ranking by choose the smallest LSR value from 10 categories
10:
9:
.
.
.
2:
1:
Reference
[1] Hani Altwaijry and Serge Belongie. Relative Ranking of Facial Attractiveness.
http://vision.ucsd.edu/sites/default/files/043-wacv.pdf

[2] Michal Uricar, Vojtech Franc and Vaclav Hlavac. Detector of Facial Landmarks Learned by the Structured Output SVM.
ftp://cmp.felk.cvut.cz/pub/cmp/articles/uricar/UricarFrancHlavac-VISAPP2012.pdf

[3] Matlab implementation of Face Detection, Pose Estimation and Landmark Localization in the Wild (FDPELL): Used to extract 68 general points of a face
http://www.ics.uci.edu/~xzhu/face/


~90.74%
Full transcript