The full text of this article/chapter:
Automatic labelling of facial features
by Juan Pablo de la Cruz Gutiérrez
12th International Workshop on Systems, Signals and Image Processing (IWSSIP), Vol. 1, No. 1, 2005
Abstract: This paper presents a first step towards the detection of facial features as well as full faces in still images. Facial features are characterized by their local visual properties. Local structure of images is extracted by projecting local image regions on a set of Gabor wavelets. Resulting coefficients are then fed into a self-organizing map (SOM) for classification, a neural network model able to automatically categorize input data extracting its most meaningful characteristics through a nonlinear and quickly converging process. The resulting mapping from feature space into the neural layer is ordered, meaning that the spatial distribution of categories responds to similarities in terms of some metric defined in the feature space. Therefore, this technique additionally provides us with insights about the statistical properties of our data. Experiments are tested on still images extracted from the audio-visual database CUAVE, in which subjects are allowed to change head pose as they speak, adding further difficulties to our approach. As starting point, we focus on the detection and localization of lips. Current results and insights on the limitations of present implementation are discussed.
is only available to individual subscribers or to users at subscribing institutions.
Pay per view: If you are not a Subscriber and you just want to read the full contents of this article, please click here to purchase online access to the full-text of this article. Please allow 3 days + mailing time. Current price for article is US$38.00
Members of the Editorial Board or subscribers of the 12th International Workshop on Systems, Signals and Image Processing (IWSSIP), that have been redirected here, please click here if you have IP-authentication access, or check if you have a registered username/password subscription with Inderscience. If that is the case, please Login: