Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SYSTEM AND METHOD FOR DETECTING FACIAL FEATURE POINT
Document Type and Number:
WIPO Patent Application WO/2017/171142
Kind Code:
A1
Abstract:
A system and a method for detecting a facial feature point are disclosed. The system for detecting a facial feature point according to an embodiment of the present invention comprises: a learning unit which divides each of a plurality of face images into regions having a predetermined size, learns a correlation between a feature vector of each divided region of the face image and the location of a predetermined feature point in the face image to thereby generate an eigen-codebook, and calculates a regression matrix indicating the correlation from the eigen-codebook; and a detection unit which detects a facial region from a target image, divides the facial region into areas having a predetermined size, detects a feature vector of each divided area of the facial region, and detects the location of a feature point of the facial region by using the detected feature vector and the regression matrix.

Inventors:
CHOI HYUN-CHUL (KR)
Application Number:
PCT/KR2016/007034
Publication Date:
October 05, 2017
Filing Date:
June 30, 2016
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
UNIV YEUNGNAM RES COOPERATION FOUNDATION (KR)
International Classes:
G06K9/00; G06T7/00
Foreign References:
KR20110067480A2011-06-22
KR101310885B12013-09-25
KR20120066462A2012-06-22
US20130044958A12013-02-21
Other References:
CHOI, HYUN-CHUL: "Latent Codebook Regression for Facial Feature Localization", ELECTRONICS LETTERS, vol. 52, no. 9, 23 February 2016 (2016-02-23), pages 1 - 2, XP006056210
CHOI, HYUN-CHU L ET AL.: "Nonparametric Facial Feature Localization Using Segment- Based Eigenfeatures", COMPUTATIONAL INTELLIGENCE AND NEUROSCIENCE, vol. 201, no. 6, 11 October 2015 (2015-10-11), pages 1 - 11, XP055430221
Attorney, Agent or Firm:
DOOHO IP LAW FIRM (KR)
Download PDF: