Method, apparatus and computer readable recording medium for detecting a location of a face feature point using an Adaboost learning algorithm

摘要

The present disclosure relates to detecting the location of a face feature point using an Adaboost learning algorithm. According to some embodiments, a method for detecting a location of a face feature point comprises: (a) a step of classifying a sub-window image into a first recommended feature point candidate image and a first non-recommended feature point candidate image using first feature patterns selected by an Adaboost learning algorithm, and generating first feature point candidate location information on the first recommended feature point candidate image; and (b) a step of re-classifying said sub-window image classified into said first non-recommended feature point candidate image, into a second recommended feature point candidate image and a second non-recommended feature point candidate image using second feature patterns selected by the Adaboost learning algorithm, and generating second feature point candidate location information on the second recommended feature point recommended candidate image.

主权项

1. A method comprising:
classifying a sub-window image into a first feature point candidate recommendation image and a first feature point candidate non-recommendation image by using first selected feature patterns, wherein a classifier includes cascaded strong classifiers, and a strong classifier in a former stage among the cascaded strong classifiers classifies the sub-window image into the first feature point candidate recommendation image and the first feature point candidate non-recommendation image using a smaller number of the first feature patterns, as compared with a strong classifier in a latter stage; generating first feature point candidate position information of the first feature point candidate recommendation image; re-classifying the sub-window image classified as the first feature point candidate non-recommendation image into a second feature point candidate recommendation image and a second feature point candidate non-recommendation image by using second selected feature patterns; generating second feature point candidate position information of the second feature point candidate recommendation image; storing the first feature point candidate position information and pieces of the second feature point candidate position information for sub-window images of a facial image; forming clusters by performing clustering in view of the first feature point candidate position information and the second feature point candidate position information; and generating position information of a certain representative position of the first feature point candidate position information and the second feature point candidate position information forming a largest cluster among the clusters as feature point position information.