Frontal View Face Research Articles

Study of effects of low level microwave field by method of face masking

The aim of this study was to examine experimentally effects of low level, modulated microwaves on human central nervous system function utilizing the phenomenon of visual masking. Ten healthy volunteers, four males and six females, were exposed to electromagnetic field (450 MHz, 0.16 mW/cm2) with 7 Hz modulation frequency. Two photo series (visual stimuli) of unfamiliar, young male faces were presented to the subjects, one picture after another. All the photos were frontal views of unfamiliar faces, which could be recognized only by their unique combinations of features. The task was to identify the pictures from a group of six photos and to decide which order they were presented in. The phenomenon of visual masking is revealed as anamorphosis in subject's perception of two instantaneous visual stimuli presented within a short time interval. When both stimuli were to be recognized correctly and put in the right order, there was a statistically significant difference (P < 0.05) between the identification of the stimulus with microwave electromagnetic field and sham exposure. Recognition of both stimuli in a pair was better under the sham exposure conditions but the actual difference was only 5%. It was concluded that early stages of visual information processing are overwhelmingly robust and routine (and adaptively significant) activities, so that the low level 7 Hz modulated electromagnetic field effects exerted upon it are extremely weak.

A new face recognition method based on SVD perturbation for single example image per person

At present, there are many methods for frontal view face recognition. However, few of them can work well when only one example image per class is available. In this paper, we present a new method based on SVD perturbation to deal with the `one example image' problem and two generalized eigenface algorithms are proposed. In the first algorithm, the original image is linearly combined with its derived image gotten by perturbing the image matrix's singular values, and then principal component analysis (PCA) is performed on the joined images. In the second algorithm, the derived images are regarded as independent images that could augment training image set, and then PCA is performed on all the training images available, including the original ones and the derived ones. The proposed algorithms are compared with both the standard eigenface algorithm and the (PC) 2A algorithm which is proposed for addressing the `one example image' problem, on the well-known FERET database with three different image resolutions. Experimental results show that the generalized eigenface algorithms are more accurate and use far fewer eigenfaces than both the standard eigenface algorithm and the (PC) 2A algorithm.

An unusual case of cranial image recognition

We encountered an unusual practical case of craniofacial identification recently. The peculiarity of the case was that the skull itself was not available for examination unlike other such common cases. The supplied material exhibits were, a nearly front view photograph of a skeletonized face and a front view face photograph of the suspected victim. Further, the condition of the skull during taking its photograph was such that its lower and the upper jaws were not in a normal closed condition. The procedure involved in dealing with such a complicated craniofacial identification problem would be quite interesting from a forensic investigator’s point of view, since standard methods of skull identification like photo/video superimposition techniques were not at all applicable here. As such, the present case report provides the details of the multiphase procedure adapted by us in dealing with this abnormal case. A solution to this unprecedented craniofacial identification problem was worked out by appropriate exploration of a newly introduced digital image processing technique that is based on craniofacial symmetry perception. The procedure leads to the reconstruction of a superimposable cranial image with upper and lower teeth in normal closed condition for establishment of its identity in usual way.

Recognition of static and dynamic images of depth-rotated human faces by pigeons.

In three experiments, we examined pigeons' recognition of video images of human faces. In Experiment 1, pigeons were trained to discriminate between frontal views of human faces in a go/no-go discrimination procedure. They then showed substantial generalization to novel views, even though human faces change radically as viewpoint changes. In Experiment 2, the pigeons tested in Experiment 1 failed to transfer to the faces dynamically rotating in depth. In Experiment 3, the pigeons trained to discriminate the dynamic stimuli showed excellent transfer to the corresponding static views, but responses to the positive faces decreased at novel viewpoints outside the range spanned by the dynamic stimuli. These results suggest that pigeons are insensitive to the three-dimensional properties of video images. Consideration is given to the nature of the task, relating to the identification of three-dimensional objects and to perceptual classifications based on similarity judgments.

Support vector machine based multi-view face detection and recognition

Detecting faces across multiple views is more challenging than in a fixed view, e.g. frontal view, owing to the significant non-linear variation caused by rotation in depth, self-occlusion and self-shadowing. To address this problem, a novel approach is presented in this paper. The view sphere is separated into several small segments. On each segment, a face detector is constructed. We explicitly estimate the pose of an image regardless of whether or not it is a face. A pose estimator is constructed using Support Vector Regression. The pose information is used to choose the appropriate face detector to determine if it is a face. With this pose-estimation based method, considerable computational efficiency is achieved. Meanwhile, the detection accuracy is also improved since each detector is constructed on a small range of views. We developed a novel algorithm for face detection by combining the Eigenface and SVM methods which performs almost as fast as the Eigenface method but with a significant improved speed. Detailed experimental results are presented in this paper including tuning the parameters of the pose estimators and face detectors, performance evaluation, and applications to video based face detection and frontal-view face recognition.

Face Detection Using a First-Order RCE Classifier

We present a new face detection algorithm based on a first-order reduced Coulomb energy (RCE) classifier. The algorithm locates frontal views of human faces at any degree of rotation and scale in complex scenes. The face candidates and their orientations are first determined by computing the Hausdorff distance between simple face abstraction models and binary test windows in an image pyramid. Then, after normalizing the energy, each face candidate is verified by two subsequent classifiers: a binary image classifier and the first-order RCE classifier. While the binary image classifier is employed as a preclassifier to discard nonfaces with minimum computational complexity, the first-order RCE classifier is used as the main face classifier for final verification. An optimal training method to construct the representative face model database is also presented. Experimental results show that the proposed algorithm yields a high detection ratio while yielding no false alarm.

Face recognition from front-view face

This letter presents a face normalization algorithm based on 2-D face model to recognize faces with variant postures from front-view face. A 2-D face mesh model can be extracted from faces with rotation to left or right and the corresponding front-view mesh model can be estimated according to the facial symmetry. Then based on the inner relationship between the two mesh models, the normalized front-view face is formed by gray level mapping. Finally, the face recognition will be finished based on Principal Component Analysis (PCA). Experiments show that better face recognition performance is achieved in this way.

Attending to faces: change detection, familiarization, and inversion effects.

We tested detection of changes to eye position, eye color (brightness), mouth position, and mouth color in frontal views of faces. Two faces were presented sequentially for 555 ms each, with a blank screen of 120 ms separating the two. Faces were presented either both upright or both inverted. Measures of detection (d') were calculated for several different degrees of change for each of the four dimensions of change. We first compared results to an earlier experiment that used an oddity design, in which subjects indicated which of three simultaneously viewed and otherwise identical faces had been altered on one of these four dimensions. Subjects in both of these experiments were partially cued, in that they knew the four possible types of changes that could occur on a given trial. The change-detection results correlated well with the oddity data. They confirmed that face inversion had little effect upon detection of changes in eye color, a moderate effect upon detection of eye-position or mouth-color changes, and caused a drastic reduction in the detection of mouth-position changes. An experiment in which uncued and fully cued subjects were compared showed that cueing significantly improved detection of feature color changes, but there was little difference between upright and inverted faces. Full cueing eliminated all effects of inversion. Compared to partial cueing, changes in mouth color were poorly detected by uncued subjects. Last, a change in the frequency of the base (unaltered) face in an experiment from 75% to 40% showed that increased short-term familiarity decreased the detection of eye changes and increased the detection of mouth changes, regardless of face orientation and the type of change made (color or position). We conclude that uncued subjects encode the spatial relations of features more than the colors of features, that mouth color in particular is not considered a relevant dimension for encoding, and that familiarization redistributes attention from more to less salient facial regions. Inversion effects are not simply an exaggeration of the salience effects revealed by withdrawing cueing, but represent an interaction of spatial encoding with salience, in that the greatest inversion effects occur for spatial shifts in less salient facial regions, and can be eliminated through the use of focused attention.

COMBINED CLASSIFICATION OF MULTIPLE VIEWS USING FACIAL CORNERS

The profile view of a face provides a complementary structure that is not seen in the frontal view. The classification system combining both frontal and profile views of faces can improve the classification accuracy. And it would be more foolproof because it is difficult to fool the profile face identification by a mask. This paper proposes a new face recognition approach, which can be applied on both frontal and profile faces, to build a robust combined multiple view face identification system. The recognition employs a novel facial corner coding and matching method, and integrates the outline and interior facial parts in the profile matching. The proposed multiview modified Hausdorff distance fuses multiple views of faces to achieve an improved system performance.

Face recognition with one training image per person

At present there are many methods that could deal well with frontal view face recognition. However, most of them cannot work well when there is only one training image per person. In this paper, an extension of the eigenface technique, i.e. projection-combined principal component analysis, (PC) 2A, is proposed. (PC) 2A combines the original face image with its horizontal and vertical projections and then performs principal component analysis on the enriched version of the image. It requires less computational cost than the standard eigenface technique and experimental results show that on a gray-level frontal view face database where each person has only one training image, (PC) 2A achieves 3–5% higher accuracy than the standard eigenface technique through using 10–15% fewer eigenfaces.

Comments on biological asymmetry

Gross forms of asymmetry of biological structure, hence of development, are generally considered abnormalities of conformation with “perfect” symmetry, whether bilateral or radial, being regarded as the “ideal” form. This notion, primarily a cultural deceit of neo-Platonic origin, can easily be shown to be wrong or at best only skin-deep by any student of anatomy or surgery who finds the heart not in the midline but, most of the time on the left, liver on the right, gut coiled and disposed in a certain direction with appendix in the right lower quadrant, and so forth. Indeed, since the beginning of Amphioxus, a major effect of evolutionary developmental modification has been the abolition of the visceral symmetry which characterized this cephalochordate with introduction of a specific pattern of asymmetry called laterality determination. This embryonic process, which is beginning to yield its universal molecular basis, is probably not responsible for another type of biological phenomenon designated fluctuating asymmetry well known to anthropologists (on the basis of quantitative studies of morphometric traits of teeth, appendicular skeleton, dermatoglyphics) and well-known to the ancients who in their most beautiful Hellenistic sculptures introduced deliberate asymmetries into facial structure and expression. Photographic images constructed of 2 right or 2 left facial halves may differ to a starling degree from the authentic face (Fig. 1). The relatively random nature of fluctuating asymmetry makes it less likely to be under strong natural selection. 1. Middle panel: Frontal view of face of a normal man. Left panel: “Artificial” face constructed out of two right halves of the same face. Right panel: Face constructed out of two left halves. A careful study of the right and left panels makes it easier to appreciate the actual degree of asymmetry present in the unaltered middle image/face. However, in addition to laterality determination and fluctuating asymmetry, there are additional forms of biological asymmetry which have other biological bases such as Lyonization, somatic/clonal mosaicism, mosaic aneuploidy/polyploidy, chimaerism, and developmental “resistance” seen with especial clarity in virtually every hereditary limb malformation. In this paper we will attempt to enumerate the causal forms and bases of biological asymmetry. © 2001 Wiley-Liss, Inc.

An economised craniofacial identification system

It has been attempted to develop an economised craniofacial identification system, as a special automated version of photo/video superimposition technique, that can deal with common cases of personal identification with the aid of a skull and a nearly front view face photograph of the suspected victim. The proposed method is economic in respect of (i) cost of hardware configuration, (ii) processing time as well as (iii) manual labour involved. Over and above, it has got a capability to take care of ambiguities due to soft tissue thickness during the selection of facial features, which is a part of the procedure. In order to reconstruct a 2-D cranial image, superimposable over the facial one, the new method does not need any reconstruction of a digitised 3-D cranial image. It works simply by a suitable segment-wise processing of a 2-D cranial image with the aid of the symmetry perceiving adaptive neuronet (SPAN), that has recently been introduced in connection with nearly front view facial image recognition. The final comparison of the facial and the superimposable cranial images is as versatile as the same for facial image recognition by SPAN. A practical application of this extended version of SPAN has been demonstrated in the present paper.

Comments on biological asymmetry

Gross forms of asymmetry of biological structure, hence of development, are generally considered abnormalities of conformation with "perfect" symmetry, whether bilateral or radial, being regarded as the "ideal" form. This notion, primarily a cultural deceit of neo-Platonic origin, can easily be shown to be wrong or at best only skin-deep by any student of anatomy or surgery who finds the heart not in the midline but, most of the time on the left, liver on the right, gut coiled and disposed in a certain direction with appendix in the right lower quadrant, and so forth. Indeed, since the beginning of Amphioxus, a major effect of evolutionary developmental modification has been the abolition of the visceral symmetry which characterized this cephalochordate with introduction of a specific pattern of asymmetry called laterality determination. This embryonic process, which is beginning to yield its universal molecular basis, is probably not responsible for another type of biological phenomenon designated fluctuating asymmetry well known to anthropologists (on the basis of quantitative studies of morphometric traits of teeth, appendicular skeleton, dermatoglyphics) and well-known to the ancients who in their most beautiful Hellenistic sculptures introduced deliberate asymmetries into facial structure and expression. Photographic images constructed of 2 right or 2 left facial halves may differ to a starling degree from the authentic face (Fig. 1). The relatively random nature of fluctuating asymmetry makes it less likely to be under strong natural selection. 1 Middle panel: Frontal view of face of a normal man. Left panel: "Artificial" face constructed out of two right halves of the same face. Right panel: Face constructed out of two left halves. A careful study of the right and left panels makes it easier to appreciate the actual degree of asymmetry present in the unaltered middle image/face. However, in addition to laterality determination and fluctuating asymmetry, there are additional forms of biological asymmetry which have other biological bases such as Lyonization, somatic/clonal mosaicism, mosaic aneuploidy/polyploidy, chimaerism, and developmental "resistance" seen with especial clarity in virtually every hereditary limb malformation. In this paper we will attempt to enumerate the causal forms and bases of biological asymmetry.

Recognizing action units for facial expression analysis

Most automatic expression analysis systems attempt to recognize a small set of prototypic expressions, such as happiness, anger, surprise, and fear. Such prototypic expressions, however, occur rather infrequently. Human emotions and intentions are more often communicated by changes in one or a few discrete facial features. In this paper, we develop an Automatic Face Analysis (AFA) system to analyze facial expressions based on both permanent facial features (brows, eyes, mouth) and transient facial features (deepening of facial furrows) in a nearly frontal-view face image sequence. The AFA system recognizes fine-grained changes in facial expression into action units (AUs) of the Facial Action Coding System (FACS), instead of a few prototypic expressions. Multistate face and facial component models are proposed for tracking and modeling the various facial features, including lips, eyes, brows, cheeks, and furrows. During tracking, detailed parametric descriptions of the facial features are extracted. With these parameters as the inputs, a group of action units (neutral expression, six upper face AUs and 10 lower face AUs) are recognized whether they occur alone or in combinations. The system has achieved average recognition rates of 96.4 percent (95.4 percent if neutral expressions are excluded) for upper face AUs and 96.7 percent (95.6 percent with neutral expressions excluded) for lower face AUs. The generalizability of the system has been tested by using independent image databases collected and FACS-coded for ground-truth by different research teams.

VIEW SYNTHESIS UNDER PERSPECTIVE PROJECTION

This paper addresses the issue of generating a 2D view of a 3D object from its other 2D views. Linear Combination method is the typical approach to this problem. However, a 2D view cannot be represented by a linear combination of other 2D views under perspective projection. Instead, we have presented a solution under perspective projection. The proposed method also applies to the construction of virtual frontal view face image and the results are encouraging.

Wavelet packet analysis for face recognition

A novel method for recognition of frontal views of human faces under roughly constant illumination is presented. The proposed scheme is based on the analysis of a wavelet packet decomposition of the face images. Each face image is first located and then, described by a subset of band filtered images containing wavelet coefficients. From these wavelet coefficients, which characterize the face texture, we build compact and meaningful feature vectors, using simple statistical measures. Then, we show how an efficient and reliable probabilistic metric derived from the Bhattacharrya distance can be used in order to classify the face feature vectors into person classes. Experimental results are presented using images from the FERET and the FACES databases. The efficiency of the proposed approach is analyzed according to the FERET evaluation procedure and by comparing our results with those obtained using the well-known Eigenfaces method.

Towards invariant face recognition

A face recognition method that can cope with the variations due to the changes in lighting and pose is presented in this paper. This system uses the output of a face detection module that extracts human faces within an input scene image. To compensate for illumination effects, an algorithm is proposed which applies an embossing operator to the detected face images to remove the illumination effects. Next, variations in the face images due to changes in pose are compensated for by using an algorithm that generates corresponding front-view face images of the detected face images; thus the perspective invariant recognition of the input face image becomes feasible. This algorithm first calculates a 3D head model from a 2D input face image. Once a 3D head is found, 2D face images can be rendered under different pose and lighting effects. For face recognition, a front-view face image is rendered. The resulting front-view face image is then directed to a fractal-based recognition algorithm. Three different experiments are carried out on a large collection of face images. The results are presented and discussed.

Effects of face configuration change on shape perception: a new illusion.

We report two experiments indicating that varying the configuration of face features changes perception of an oval aperture windowing the face: as the eyes and mouth of a frontal-view face photograph are moved vertically toward face boundaries, the oval appears increasingly clongated, taller, and narrower; when eyes and mouth are moved toward the nose, the oval appears increasingly rounder, shorter, and wider. This shape illusion is maximised when faces appear upright within the oval and major face features (eyes, nose, and mouth) appear in their correct relative locations. These results establish that processing of a face configuration can affect perception of a geometric shape that shares visual space with a face. Whether the illusion is face-specific or a special case of a more general geometric illusion is discussed.

Frontal-view face detection and facial feature extraction using color and morphological operations

A novel algorithm for front-view facial contour detection and features extraction is described. A skin-color face segmentation method is developed to detect the face region firstly. In order to more precisely locate the face contour and features such as eyes, mouth and nostrils, we eliminate the ears and neck from the face region by using morphological operations and knowledge about the face structure. We have found some consistent rules that can be used to locate some contour points from the skeleton of the processed frontal face region. The face contour can then be fitted as an ellipse using these points. Then the facial features can be located in the interior of the face contour. Experiments have been done with a number of images of frontal view of the head including some with a slight pan obtained from an Internet face database (University of Stirling). Also a number of face images captured by a color CCD camera are tried by our method. The robustness is confirmed with correct detection rate of over 90%.

Viewpoint-dependent recognition of familiar faces.

The question whether object representations in the human brain are object-centered or viewer-centered has motivated a variety of experiments with divergent results. A key issue concerns the visual recognition of objects seen from novel views. If recognition performance depends on whether a particular view has been seen before, it can be interpreted as evidence for a viewer-centered representation. Earlier experiments used unfamiliar objects to provide the experimenter with complete control over the observer's previous experience with the object. In this study, we tested whether human recognition shows viewpoint dependence for the highly familiar faces of well-known colleagues and for the observer's own face. We found that observers are poorer at recognizing their own profile, whereas there is no difference in response time between frontal and profile views of other faces. This result shows that extensive experience and familiarity with one's own face is not sufficient to produce viewpoint invariance. Our result provides strong evidence for viewer-centered representations in human visual recognition even for highly familiar objects.