R. Baron, "Mechanisms of human facial recognition," Int. J. Man--Machine Studies 15, 137--178 (1981).  Home/Search   Document Not in Database   Summary   Related Articles   Check

....surveillance, visually mediated interaction, human machine interface, multimedia and teleconferencing. Various approaches have been proposed to address the problem under different assumptions and conditions. Template based methods were adopted in many previous studies. Early work by Baron [1] presented a Neural Network based approach to face recognition using raw images as sys tem input. Recognition was performed based on the correlation of the resulting sequence of patterns with all model patterns. Brunelli and Poggio [5] presented and compared a ge ometrical feature based ....

R. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137-178, 1981.

....measure a total of 35 geometric features of a face . First, a set of 5 eye templates is constructed by scaling the eyes of one of the authors (the set of scales used is 0.7, 0.85, 1, 1.15, 1. 3 to account for the expected scale variation ) Then, a cross correlation process similar to that of [21] is performed to locate the eyes. Once the eyes are located, they take the advantage of the knowledge of anthropometric standards to search for the other features, including mouth, nose, eyebrow, and face outline. To speak in more detail, mouth and nose are located using the technique of ....

....practical system should be able to handle these variations. In [22] Brunelli and Poggio made extensive experiments on a geometric feature approach (described in in Section 2.3.1) and a template matching approach to face recognition. Their template matching approach is based on the work of Baron [21]. First, the images is normalized using the same technique described in Section 2.3.1. Each person is represented by a database entry whose fields are a digital image of his her frontal view and a set of four masks representing eyes, nose, mouth, and face. In their implementation, the matching ....

R. J. Baron, "Mechanism of human facial recognition," International Journal of Man Machine Studies, 15:137--178, 1981.

....glasses left light (10) sun glasses right light, 11) scarf, 12) scarf left light, 13) scarf right light 3 Face Recognition Algorithms Most of the current face recognition algorithms can be categorized into two classes, image template based or geometry feature based. The template based methods [1] compute the correlation between a face and one or more model templates to estimate the face identity. Statistical tools such as Support Vector Machines (SVM) 30, 20] Linear Discriminant Analysis (LDA) 2] Principal Component Analysis (PCA) 28, 29] Kernel Methods [26, 16] and Neural Networks ....

Robert J. Baron. Mechanisms of human facial recognition. International Journal of Man-Machine Studies, 15(2):137--178, 1981.

....distinctive parameters like eyes, nose, mouth and chin even in pictures of very coarse resolution. The other important technical approach to face recognition is using template matching. In this approach, templates of a face are used to match to that in the picture with varying parameters as in [2]. In [9] the two approaches are compared in terms of results and performance. Next, wewillgive a brief description of a technique named snakes that can be employed in template matching. This technique could be incorporated into our system for improving the feature extraction process. Various ....

R. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....techniques are usually needed. We can do tem plate matching after preprocessing the images. We can also use multiple templates to take into account multiple viewpoints etc. We can even use multiple templates for each viewpoint each covering a smaller area of the face. See for example Baron ([11]) and Bischel ( 23] 1i.e. the presence of a face may be detected 126 Deformable Templates A more complex approach is to use a deformable template, where we allow the tem plate to be deformed before matching; but penalizing the deformation by putting a cost to it. This is equivalent to ....

R.J. Baron, "Mechanisms of Human Facial Recognition" International Journal of Man Machine Studies vol 15, p. 137 178, 1981

....with error rates of less than 10 percent. These successes have led to the development of a number of commercial face recognition systems. Most of the current face recognition algorithms can be categorized into two classes, image template based or geometry feature based. The template based methods [1] compute the correlation between a face and one or more model templates to estimate the face identity. Statistical tools such as Support Vector Machines (SVM) 30, 21] Linear Discriminant Analysis (LDA) 2] Principal Component Analysis (PCA) 27, 29, 11] Kernel Methods [25, 17] and Neural ....

Robert J. Baron. Mechanisms of human facial recognition. International Journal of Man-Machine Studies, 15(2):137--178, 1981.

....in [21] 22] Most of this work is either recognition by using facial profiles (for example, see [23] 24] or using the frontal views. In this paper we are interested in the latter case where the input is an intensity image of the frontal view of a face. Previous related work can be found in [25], the WISARD system ( 26] 27] and the dynamic link architecture for face recognition [28] One of the early systems built for this task is described in [29] The system automatically localizes features such as corners of the eyes, nostrils, mouth etc. Then a set of sixteen facial parameters ....

R. J. Baron, "Mechanisms of human facial recognition," Intl. Journal of Man-Machine Studies, vol. 15, pp. 137--178, 1981.

....with error rates of less than 10 percent. These successes have led to the development of a number of commercial face recognition systems. Most of the current face recognition algorithms can be categorized into two classes, image template based or geometry feature based. The template based methods [1] compute the correlation between a face and one or more model templates to estimate the face identity. Statistical tools such as Support Vector Machines (SVM) 30, 21] Linear Discriminant Analysis (LDA) 2] Principal Component Analysis (PCA) 27, 29, 11] Kernel Methods [25, 17] and Neural ....

Robert J. Baron. Mechanisms of human facial recognition. International Journal of Man-Machine Studies,

....Subsequent symmetries lie within features such as the eyes, nose and mouth. The symmetry operator locates points in the image corresponding to high values of a symmetry measure discussed in detail in [47] The procedure is claimed to be superior to other correlation based schemes such as that of [48] in the sense that it is independent of scale or orientation. However, since no a priori knowledge of face location is used, the search for symmetry points is computationally intensive. A success rate of 95 is reported on a face image database, with the constraint that the faces occupy between ....

R. Baron, "Mechanisms of Human Facial Recognition," International Journal of Man-Machine Studies, Vol. 15, pp. 137--178, 1981.

....Also, 3 D motion parameters have to be computed precisely from image data. 2. 2 Work on face images and recognition A good survey of the state of the art on image processing of faces is given in [12] However, most of the work with faces in computer vision was done on processing for recognition [33, 6, 61, 36, 39, 3, 2, 4, 15, 28, 9] and some work on different kinds of classification tasks [21, 22, 14] Almost all of this work treats face recognition as a static problem approached by pattern recognition techniques 1 applied to single static images. Only recently the attention of researchers has shifted to the temporal aspect ....

....some ideas that are relevant to our work. Recently, a systematic comparison of typical approaches (feature based versus template based techniques) to face recognition was carried out by Brunelli Poggio [13, 15] Several other approaches to face recognition have also been presented (for example [6, 61, 36, 4]) The first approach is influenced by the work of Kanade [33] and uses a vector of geometrical features for recognition. First the eyes are located by computing the normalized cross correlation coefficient with a single eye template at different resolution scales. The image is then normalized in ....

[Article contains additional citation context not shown here]

R. J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....facial recognition [Hay and Young, 1982] The most difficult faces for humans to recognize are those faces which are considered neither attractive nor unattractive by the observer. This gives support to the theories suggesting that distinctive faces are more easily recognized than typical ones [Baron, 1981]. Information contained in low spatial frequency bands is used in order to make the determination of the sex of the individual, 4 while the higher frequency components are used in recognition [Sergent, 1986] Young children typically recognize unfamiliar faces using unrelated cues, such as ....

Baron, R. (1981). Mechanisms of human facial recognition. International Journal of Man-Machine Studies, pages 137--178.

....utilisees pour identifier les visages sont alors obtenues par des calculs de correlation entre les regions. D autres methodes utilisent une analyse en composantes principales realisee a partir de l information pixel, pour rechercher une base de representation permettant de discriminer les visages [1] [2] 15] Le systeme de reconnaissance de visage, que nous presentons, s inscrit dans le cadre du projet AMIBE, un des projets de recherche du PRC Communication Homme Machine, qui se propose d experimenter une interface multimodale homme machine integrant le son et l image pour un nombre limite ....

R. J. Baron. Mechanisms of human facial recognition. The International Journal of Robotics Research, 15:137, 1981.

....example views per person. We will then discuss work on synthesizing virtual example views to deal with situations where only one example view (also called model view) per person is available. 4.2. 1 The Brunelli Poggio frontal face recognition system Following the face recognition work of Baron [1], Brunelli and Poggio [9] use a templatebased strategy to recognize frontal views of faces. From an example model view, faces are represented using templates of the eyes, nose, mouth, and entire face. A normalized correlation metric is used to compare the model templates against the input image. ....

Robert J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....used in the context of image based face recognition techniques namely, template based and neural networks. In the template based approach, the face is represented as a set of templates of the major facial features which are then matched with the prototypical model face templates (see, for example, [2]) Extensions to this technique include low dimensional coding to simplify the template representation and improve the performance of the template matching process (see, for example, the eigenfaces of Turk and Pentland [9] stochastic modeling with Hidden Markov Models (HMMs) 8] and, ....

R. Baron. Mechanisms of human facial recognition. Int. Journal Man Machine Studies, 15:137--178, 1981.

....[8] developed procedures to locate human faces in newspaper photographs. Their approach is based on cost minimization of feature graphs. Turk and Pentland [16] find the face by analyzing frame differencing in motion under the hypothesis that people are constantly moving. In an early paper, Baron [3] shows how to locate the eyes in a face. His approach is based on the standardization of the image size. Yuille, Cohen and Hallinan [19] use deformable templates and an energy minimization function to find the eyes. Ohmura, Tomono, and Kobayashi presented a method of detection of face direction ....

Robert J. Baron. Mechanisms of Human Facial Recognition. Int. J. Man-Machine Studies, 15:137--178, 1981.

....found in the literature from which we can mention here only a few. A good and more comprehensive survey of the state of the art can be found in [2, 3] Early algorithms [4] use feature based techniques (e.g. features like localization or thickness of eyebrows) or template matching for recognition [5]. Recently, a systematic comparision of feature based vs. template based algorithms has been undertaken by Brunelli and Poggio [6] Gilbert and Yang [7] presented a real time face recognition system using custom VLSI hardware. The hardware allows fast template correlation and the system is able to ....

R. J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....facial features such as the eye shape, nose location, and cheek bone curvature. These few extracted facial parameters are subsequently compared to database of known faces. Parameter based recognition schemes attempt to develop an efficient representation of salient features of an individual [1]. While the database search and comparison for parameter based recognition may not be computationally intensive, the image processing required to extract the appropriate parameters is quite computationally expensive and requires careful selection of facial parameters which will unambiguously ....

Robert J. Baron, "Mechanisms of human facial recognition," International Journal of Man-Machine Studies, vol. 15, pp. 137-178, 1981.

....we will illustrate the properties of a set of HyperBF networks that use geometrical features as inputs and are trained to identify people. As we discuss in a separate paper [8] it seems that better performance in this specific task can be achieved by schemes (see for instance [14] 15] 8] [16] that use templates rather than geometric features. HyperBF networks could also accept as inputs pixel values instead than geometrical features but it is unlikely that they will perform better than traditional classifiers for these inputs. It is not impossible, however, that geometric like ....

R. J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....of the face and nose, nostrils) and then capture the spatial configuration in feature vector whose dimensions typically include measurements like distances, angles, and curvatures. Pictorial approaches, representing faces by using filtered images of model faces, include template based systems ([2], 6] 13] 7] and [5] systems using principal components analysis to derive a pictorial face space ( 15] 20] 1] 9] and connectionist approaches ( 16] 11] 10] 21] and [12] 18] explores an interesting hybrid representation that combines the geometrical and pictorial ....

....and flexible matching strategies [18] to handle some degree of expression and out of plane rotations. What distinguishes our approach from these techniques will be a wider allowed variation in viewpoint. Overall, while face recognition systems have been successful (the template based systems in [2] and [6] achieved 100 recognition on a data base of over 40 people) most recognition systems work with frontal views, no expressions, and controlled lighting. Our goal is to build a face recognizer that works under varying pose, the difficult part of which is to handle face rotations in depth. ....

[Article contains additional citation context not shown here]

Robert J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....N00014 91 J 4038. D. Beymer is supported by a Howard Hughes Doctoral Fellowship from the Hughes Aircraft Company. 1 Introduction Existing work in face recognition has demonstrated good recognition performance on frontal, expressionless views of faces with controlled lighting (see Baron [4], Turk and Pentland [48] Bichsel [11] Brunelli and Poggio [14] and Gilbert and Yang [20] One of the key remaining problems in face recognition is to handle the variability in appearance due to changes in pose, expression, and lighting conditions. There has been some recent work in this ....

Robert J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

....which avoids the difficulty that we do not understand the recognition process sufficiently well to be able to write a recognition program; and they can meet a real time requirement. 8 Baron firstly proposed the visual perception and recognition of faces in terms of the underlying neural networks [39]. He suggested several different networks and mechanisms that underlie face recognition. He described the logical operations that underlie face recognition and proposed neural networks to perform the operations. The first stage of deriving memory representations requires that the input images be ....

....poses of photographs, i.e. front face with neutral expression; face turning left 20 ffi ; face turning right 20 ffi ; front face with smile; and front face with some form of disguise (taking off or wearing glasses) Only the first three poses were tested. This model made use of Baron s idea [39] of using different resolution between the whole face image and detailed regions. A multilayer perceptron network was developed to recognize human faces independent of translation, rotation and perspective changes [47] The network forms its own set of face features in the hidden layer. The ....

R.J. Baron. Mechanisms of human facial recognition. Int. J. Man - Machine Studies, 15:137--178, 1981.

....preprocessed before matching. Several full templates per each face may be used to account for the recognition from different viewpoints. Still another important variation is to use, even for a single viewpoint, multiple templates. A face is stored then as a set of distinct(ive) smaller templates [1]. A rather different approach is based on the technique of elastic templates [6] 5] 23] 2. Experimental setup The database we used for the comparison of the different strategies is composed of 188 images, four for each of 47 people. Of the four pictures available, the first two were taken in ....

....followed achieves scale and rotation invariance by setting the interocular distance and the direction of the eye to eye axis. We will describe the steps of the normalization procedure in some detail since they are themselves of some interest. The first step in our technique resembles that of Baron [1] and is based on template matching by means of a normalized cross correlation coefficient, defined by : CN (y) IT T Gamma IT T oe(IT )oe(T ) 1) where IT be the patch of image I which must be matched to T , the average operator, IT T represent the pixel by pixel product, and oe ....

R. J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

.... processes used by people: iconic: it is based on the comparison of suitably preprocessed image patches; recognition is effected by comparing (e.g. through the value of cross correlation or some other suitable distance measure) an unknown image with stored templates of distinctive facial regions [3], 25] 21] 22] 6] geometric: a set of geometric features, describing the size and the layout of the different features in the faces, is computed and recognition proceeds by comparing the unknown descriptive vector with a set of reference vectors (known people) stored in a data base [13] ....

R. J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

No context found.

R. Baron, "Mechanisms of human facial recognition," Int. J. Man--Machine Studies 15, 137--178 (1981).

No context found.

R. J. Baron. Mechanisms of human facial recognition. International Journal of Man Machine Studies, 15:137--178, 1981.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC