METHOD FOR DETECTING FACIAL FEATURES

• US 20120328199A1
• Filed: 08/22/2011
• Published: 12/27/2012
• Est. Priority Date: 06/24/2011
• Status: Active Grant

• Current Assignee: LG Innotek Company Limited (LG Corporation)
• Original Assignee: LG Innotek Company Limited (LG Corporation)
• Inventors: Woo, Jeong Woo, Talluri, Mallikarjuna Rao
• Granted Patent: US 8,687,855 B2
• Time in Patent Office: Days
• US Class Current: 382/197
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis
• CPC Class Codes: G06V 40/165 using facial parts and geometric relationships

Abstract

A method for detecting facial feature is provided. Concerned points and concerned areas are extracted from a captured image, and an area which is considered to highly possibly include a facial area is selected. It is determined whether the selected area corresponds to a facial feature by using a support vector machine.

Claims

1. A method for detecting a facial feature, the method comprising:

extracting a face area from a captured image;

extracting concerned points from the extracted face area;

dividing the extracted face area to a plurality of sub areas;

extracting local maxima points from the plurality of sub areas;

selecting optimum areas, which have highest similarities with a feature image, around the extracted local maxima points; and

determining which one of the selected optimum areas has a largest value of similarity among the highest similarities.

2. The method according to claim 1, wherein the extracting of the face area is performed by using Adaboost algorithm.

3. The method according to claim 1, wherein the extracting of the concerned points is performed by using radial symmetry transform.

4. The method according to claim 1, wherein the selecting of the optimum areas comprises:

configuring a mask having a predetermined size around the local maxima point;

measuring similarities between a facial feature and areas obtained by shifting the mask vertically and horizontally as much as a predetermined distance; and

detecting an area which has highest similarity by moving the mask as much as the predetermined distance in a direction yielding higher similarity.

5. The method according to claim 1, wherein the similarity with the facial feature is measured by a support vector machine.