A method for automatically classifying facial images is proposed. Faces are represented using elastic graphs labelled with with 2D Gabor wavelet features. The system is trained from examples to classify faces on the basis of high-level attributes, such as sex, "race", and expression, using linear discriminant analysis (LDA). Use of the Gabor representation relaxes the requirement for precise normalization of the face: approximate registration of a facial graph is sufficient. LDA allows simple and rapid training from examples, as well as straightforward interpretation of the role of the input features for classification. The algorithm is tested on three different facial image datasets, one of which was acquired under relatively uncontrolled conditions, on tasks of sex, "race" and expression classification. Results of these tests are presented. The discriminant vectors may be interpreted in terms of the saliency of the input features for the different classification tasks, which we portray visually with feature saliency maps for node position as well as filter spatial frequency and orientation.

A method of manifold representation for human faces with pose variations is proposed. Our model consists of mappings between 3D head angles and facial images separately represented in shape and texture, via sub-space models spanned by principal components (PC). Explicit mappings to and from 3D head angles are used as processes of pose estimation and transformation, respectively. Generalization capability to unknown head poses enables our model to continuously cover pose parameter space, providing high approximation accuracy. The feasibility of this model is evaluated in a number of experiments. We also propose a novel pose-invariant face recognition system using our model as the entry format for a gallery of known persons. Experimental results with 3D facial models recorded by a Cyberware scanner show that our model provides a superior recognition performance against pose variations, and that the texture synthesis process is carried out correctly.

This paper presents an analysis of the performance of two different skin chrominance models and of nine different chrominance spaces for the color segmentation and subsequent detection of human faces in two-dimensional static images. For each space, we use the single Gaussian model based on the Mahalanobis metric and a Gaussian mixture density model to segment faces from scene backgrounds. In the case of the mixture density model, the skin chrominance distribution is estimated by use of the expectation-maximisation (EM) algorithm. Feature extraction is performed on the segmented images by use of invariant Fourier-Mellin moments. A multilayer perceptron neural network (NN), with the invariant moments as the input vector, is then applied to distinguish faces from distractors. With the single Gaussian model, normalized color spaces are shown to produce the best segmentation results, and subsequently the highest rate of face detection. The results are comparable to those obtained with the more sophisticated mixture density model. However, the mixture density model improves the segmentation and face detection results significantly for most of the un-normalized color spaces. Ultimately, we show that, for each chrominance space, the detection efficiency depends on the capacity of each model to estimate the skin chrominance distribution and, most importantly, on the discriminability between skin and "non-skin" distributions.

We apply the orthogonal Fourier-Mellin moments (OFMMs) to the specific problem of fully translation-, scale- and in-plane rotation-invariant detection of human faces in 2D static color images, and compare their performance with that of the generalized Hu's moments or non-orthogonal FMMs. The OFMMs have the advantages of non-redundancy of information, robustness with respect to noise and the ability to reconstruct the original object. Color segmentation is first performed in nine different chrominance spaces by use of two human skin chrominance models. For feature extraction in the segmented images, the same number of OFMMs are used (as for the FMMs) as the input vector to a multilayer perceptron neural network to distinguish faces from distracters. It is shown that, at least in the specific problem of face detection from segmented images, for the same set of test images, there is no significant advantage over the FMMs in using the OFMMs, and that in practice both types of moments may be used. Possible explanations for such results are presented.

This paper present an analysis of the performance of support vector machines (SVMs) for automatic detection of human faces in static color images of complex scenes. Skin color-based image segmentations initially performed for several different chrominance spaces by use of the single Gaussian chrominance model and a Gaussian mixture density model. Feature extraction in the segmented images is then implemented by use of invariant orthogonal Fourier-Mellin moments. For all chrominance spaces, the application of SVMs to the invariant moments obtained from a set of 100 test images yields a higher face detection performance than when applying a 3-layer perceptron neural network (NN), depending on a suitable selection of the kernel function used to train the SVM and of the value of its associated parameter(s). The training of SVMs is easier and faster than that of a NN, always finds a global minimum, and SVMs have a better generalization ability.

A method for automatically extracting symbolic-level information from single facial images is proposed. We combine a 2D Gabor wavelet labelled elastic graph representation with Fisher discriminant analysis. Use of this representation relaxes the requirement for exact normalization of the face. The algorithm is used to extract information about the sex, "race" and emotional expression of the face in the image. A visual interpretation of the discriminant vectors is provided.

We describe results obtained from a testbed used to investigate different codings for automatic face recognition. An eigenface coding of shape-free faces using manually located landmarks was more effective than the corresponding coding of correctly shaped faces. Configuration also proved an effective method of recognition, with rankings given to incorrect matches relatively uncorrelated with those from shape-free faces. Both sets of information combine to improve significantly the performance of either system. The addition of a system, which directly correlated the intensity values of shape-free images, also significantly increased recognition, suggesting extra information was still available. The recognition advantage for shape-free faces reflected and depended upon high-quality representation of the natural facial variation via a disjoint ensemble of shape-free faces; if the ensemble comprised nonfaces, a shape-free disadvantage was induced. Manipulation within the shape-free coding to emphasize distinctive features of the faces, by caricaturing, allowed further increases in performance; this effect was only noticeable when the independent shape-free and configuration coding was used. Taken together, these results strongly support the suggestion that faces should be considered as lying in a high-dimensional manifold, which is locally linearly approximated by these shapes and textures, possibly with a separate system for local features. Principal components analysis is then seen as a convenient tool in this local approximation.

We propose a method for automatically classifying facial images based on labeled elastic graph matching, a 2D Gabor wavelet representation, and linear discriminant analysis. Results of tests with three image sets are presented for the classification of sex, "race", and expression. A visual interpretation of the discriminant vectors is provided.

Use of uncalibrated images has found many applications such as image synthesis. However, it is not easy to specify the desired position of the new image in projective or affine space. This paper proposes to recover Euclidean structure from uncalibrated images using domain knowledge such as distances and angles. The knowledge we have is usually about an object category, but not very precise for the particular object being considered. The variation (fuzziness) is modeled as a Gaussian variable. Six types of common knowledge are formulated. Once we have an Euclidean description, the task to specify the desired position in Euclidean space becomes trivial. The proposed technique is then applied to synthesis of new facial images. A number of difficulties existing in image synthesis are identified and solved. For example, we propose to use edge points to deal with occlusion.

We use a skin color model based on the Mahalanobis metric and a shape analysis based on invariant Fourier-Mellin moments to automatically detect and locate human faces in two-dimensional complex scene images. First, color segmentation of an input image is performed by thresholding in a normalized hue-saturation. Color space where the effects of the variability of human skin color and the dependency of chrominance on changes in illumination are reduced. We then group regions of the resulting binary image that have been classified as face candidates into clusters of connected pixels. Discarding the smallest clusters in the image ensures that only a small number of clusters will be used for further analysis. Fully translation-, scale- and in-plane rotation-invariant moments are calculated for each remaining cluster Finally, in order to distinguish faces from distractors, a multilayer perceptron neural network is used with the invariant moments as the input vector supervised learning of the network is implemented with the backpropagation algorithm, at first for frontal views of faces. Preliminary results show the efficiency of the combination of color segmentation and of invariant moments in detecting faces with a large variety of poses and against relatively complex backgrounds.

Article Avatar creation using automatic face processing Share on Authors: Michael Lyons ATR Human Information Processing Research Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan ATR Human Information Processing Research Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, JapanView Profile , Andre Plante ATR Media Integration and Communication Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan ATR Media Integration and Communication Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, JapanView Profile , Sebastien Jehan ATR Human Information Processing Research Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan ATR Human Information Processing Research Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, JapanView Profile , Seiki Inoue ATR Media Integration and Communication Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan ATR Media Integration and Communication Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, JapanView Profile , Shigeru Akamatsu ATR Human Information Processing Research Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, Japan ATR Human Information Processing Research Laboratory, 2-2 Hikaridai, Seika-cho, Soraku-gun, Kyoto 619-02, JapanView Profile Authors Info & Claims MULTIMEDIA '98: Proceedings of the sixth ACM international conference on MultimediaSeptember 1998 Pages 427–434https://doi.org/10.1145/290747.290816Online:01 September 1998Publication History 17citation614DownloadsMetricsTotal Citations17Total Downloads614Last 12 Months20Last 6 weeks2 Get Citation AlertsNew Citation Alert added!This alert has been successfully added and will be sent to:You will be notified whenever a record that you have chosen has been cited.To manage your alert preferences, click on the button below.Manage my AlertsNew Citation Alert!Please log in to your account Save to BinderSave to BinderCreate a New BinderNameCancelCreateExport CitationPublisher SiteGet Access

We derive a machine-based face recognition system theoretically which is similar to many practical ones and show its behaviour has features of the human system. We reproduce the caricature effect, show that human and machine-based similarity and distinctiveness are connected and confirm machine based predictions of typicality with human data. Finally we draw lessons for machine-based recognition.

An effective algorithm for character recognition in scene images is studied. Scene images are segmented into regions by an image segmentation method based on adaptive thresholding. Character candidate regions are detected by observing gray-level differences between adjacent regions. To ensure extraction of multisegment characters as well as single-segment characters, character pattern candidates are obtained by associating the detected regions according to their positions and gray levels. A character recognition process selects patterns with high similarities by calculating the similarities between character pattern candidates and the standard patterns in a dictionary and then comparing the similarities to the thresholds. A relaxational approach to determine character patterns updates the similarities by evaluating the interactions between categories of patterns, and finally character patterns and their recognition results are obtained. Highly promising experimental results have been obtained using the method on 100 images involving characters of different sizes and formats under uncontrolled lighting.>

Proposes a scheme that offers accurate and robust identification of human faces. The scheme is characterized by four aspects: facial feature detection using color image segmentation; target image extraction using a sub-space classification method; robust feature extraction based on K-L expansion of an invariant feature space; and face classifier training based on 3D CG modeling of the human face. The scheme's flexibility under a wide range of image acquisition conditions has been confirmed through the assessment of an experimental face identification system.>

Proposes a variable filter, a new high-resolution Hough transform method. A voted value at the point ( rho , theta ) on the Hough space is transformed into a value distribution by the variable filter, and voting is performed according to the distribution. The shape and size of the variable filter changes dynamically with the quantization size and the line parameters ( rho , theta ). Statistical experiments show that the filter can improve extraction performance and reduce the mean error of extracted line parameters even with noisy images.>

An effective extraction algorithm for character recognition in scene images is developed. Character candidate regions are detected by an image segmentation method based on adaptive thresholding and by evaluating gray-level difference between adjacent regions. Character pattern candidates are obtained by associating the detected regions according to their positions and gray levels and evaluating the aspect ratio of the rectangle that circumscribes the associated regions. The character recognition process selects character pattern candidates with high similarities to any categories in a dictionary (high similarity patterns). Highly promising experimental results have been obtained using the method on 100 images acquired under uncontrolled lighting. This algorithm can be applied to read characters of any size and format under various lighting conditions.>