Showing papers on "Facial recognition system published in 1993"

High confidence visual recognition of persons by a test of statistical independence

Abstract: A method for rapid visual recognition of personal identity is described, based on the failure of a statistical test of independence. The most unique phenotypic feature visible in a person's face is the detailed texture of each eye's iris. The visible texture of a person's iris in a real-time video image is encoded into a compact sequence of multi-scale quadrature 2-D Gabor wavelet coefficients, whose most-significant bits comprise a 256-byte "iris code". Statistical decision theory generates identification decisions from Exclusive-OR comparisons of complete iris codes at the rate of 4000 per second, including calculation of decision confidence levels. The distributions observed empirically in such comparisons imply a theoretical "cross-over" error rate of one in 131000 when a decision criterion is adopted that would equalize the false accept and false reject error rates. In the typical recognition case, given the mean observed degree of iris code agreement, the decision confidence levels correspond formally to a conditional false accept probability of one in about 10/sup 31/. >

Face recognition: features versus templates

Abstract: Two new algorithms for computer recognition of human faces, one based on the computation of a set of geometrical features, such as nose width and length, mouth position, and chin shape, and the second based on almost-gray-level template matching, are presented. The results obtained for the testing sets show about 90% correct recognition using geometrical features and perfect recognition using template matching. >

Distortion invariant object recognition in the dynamic link architecture

Abstract: An object recognition system based on the dynamic link architecture, an extension to classical artificial neural networks (ANNs), is presented. The dynamic link architecture exploits correlations in the fine-scale temporal structure of cellular signals to group neurons dynamically into higher-order entities. These entities represent a rich structure and can code for high-level objects. To demonstrate the capabilities of the dynamic link architecture, a program was implemented that can recognize human faces and other objects from video images. Memorized objects are represented by sparse graphs, whose vertices are labeled by a multiresolution description in terms of a local power spectrum, and whose edges are labeled by geometrical distance vectors. Object recognition can be formulated as elastic graph matching, which is performed here by stochastic optimization of a matching cost function. The implementation on a transputer network achieved recognition of human faces and office objects from gray-level camera images. The performance of the program is evaluated by a statistical analysis of recognition results from a portrait gallery comprising images of 87 persons. >

Masking the face: Recognition of emotional facial expressions as a function of the parameters of backward masking

Abstract: Four experiments are reported investigating recognition of emotional expressions in very briefly presented facial stimulus. The faces were backwardly masked by neutral facial displays and recognition of facial expressions was analyzed as a function of the manipulation of different parameters in the masking procedure. The main conclusion was that stimulus onset asynchrony between target and mask proved to be the principal factor influencing recognition of the masked expressions. In general, confident recognitions of facial expressions required about 100-150 msec, with shorter time for happy than for angry expressions. The manipulation of the duration of both the target and the mask, by itself, had only minimal effects.

Low-dimensional representation of faces in higher dimensions of the face space

Abstract: Faces can be represented efficiently as a weighted linear combination of the eigenvectors of a covariance matrix of face images. It has also been shown [ J. Opt. Soc. Am.4, 519– 524 ( 1987)] that identifiable faces can be made by using only a subset of the eigenvectors, i.e., those with the largest eigenvalues. This low-dimensional representation is optimal in that it minimizes the squared error between the representation of the face image and the original face image. The present study demonstrates that, whereas this low-dimensional representation is optimal for identifying the physical categories of face, like sex, it is not optimal for recognizing the faces (i.e., discriminating known from unknown faces). Various low-dimensional representations of the faces in the higher dimensions of the face space (i.e., the eigenvectors with smaller eigenvalues) provide better information for face recognition.

Automatic creation of 3D facial models

Abstract: Model-based encoding of human facial features for narrowband visual communication is described. Based on an already prepared 3D human model, this coding method detects and understands a person's body motion and facial expressions. It expresses the essential information as compact codes and transmits it. At the receiving end, this code becomes the basis for modifying the 3D model of the person and thereby generating lifelike human images. The feature extraction used by the system to acquire data for regions or edges that express the eyes, nose, mouth, and outlines of the face and hair is discussed. The way in which the system creates a 3D model of the person by using the features extracted in the first part to modify a generic head model is also discussed. >

Optical network for real-time face recognition

Abstract: An optical network is described that is capable of recognizing at standard video rates the identity of faces for which it has been trained. The faces are presented under a wide variety of conditions to the system and the classification performance is measured. The system is trained by gradually adapting photorefractive holograms.

Multi-modular neural network architectures: applications in optical character and human face recognition

Abstract: In practical applications, recognition accuracy is sometimes not the only criterion; capability to reject erroneous patterns might also be needed. We show that there is a trade-off between these two properties. An efficient solution to this trade-off is brought about by the use of different algorithms implemented in various modules, i.e. multi-modular architectures. We present a general mechanism for designing and training multi-modular architectures, integrating various neural networks into a unique pattern recognition system, which is globally trained. It is possible to realize, within the system, feature extraction and recognition in successive modules which are cooperatively trained. We discuss various rejection criteria for neural networks and multi-modular architectures. We then give two examples of such systems, study their rejection capabilities and show how to use them for segmentation. In handwritten optical character recognition, our system achieves performances at state-of-the-art level, but is eight times faster. In human face recognition, our system is intended to work in the real world.

Identification of human faces through texture-based feature recognition and neural network technology

Abstract: A method is presented to infer the presence of a human face in an image through the identification of face-like textures. The selected textures are those of human hair and skin. The second-order statistics method is used for texture representation. This method employs a set of co-occurrence matrices, from which features can be calculated that can characterize a texture. The cascade-correlation neural network architecture is used for supervised classification of textures. The Kohonen self-organizing feature map shows the clustering of the different texture types. Classification performance is generally above 80%, which is sufficient to clearly outline a face in an image. >

A real-time face recognition system using custom VLSI hardware

Abstract: A real-time face recognition system has been implemented on an IBM compatible personal computer with a video camera, image digitizer, and custom VLSI image correlator chip. With a single frontal facial image under semicontrolled lighting conditions, the system performs (i) image preprocessing and template extraction, (ii) template correlation with a database of 173 images, and (iii) postprocessing of correlation results to identify the user. System performance issues including image preprocessing, face recognition algorithm, software development, and VLSI hardware implementation are addressed. In particular, the parallel, fully pipelined VLSI image correlator is able to perform 340 Mop/second and achieve a speed up of 20 over optimized assembly code on a 80486/66DX2. The complete system is able to identify a user from a database of 173 images of 34 persons in approximately two to three seconds. While the recognition performance of the system is difficult to quantify simply, the system achieves a very conservative 88% recognition rate using cross-validation on the moderately varied database.

Eyes detection for face recognition

Abstract: A correlation-based approach to automatic face recognition requires adequate normalization techniques. If the positioning of the face in the image is accurate, the need for shifting to obtain the best matching between the unknown subject and a template is drastically reduced, with considerable advantages in computing costs. In this paper, a novel technique is presented based on a very efficient eyes localization algorithm. The technique has been implemented as part of the “electronic librarian” of MAIA, the experimental platform of the integrated Al project under development at IRST. Preliminary experimental results on a set of 220 facial images of 55 people disclose excellent recognition rates and processing speed.

Study on face robot for active human interface-mechanisms of face robot and expression of 6 basic facial expressions

Abstract: In order to develop an active human interface (AHI) that realizes heart-to-heart communication between intelligent machine and human being, we have been undertaking the investigation of the method for improving the sensitivity of "KANSEI" communication between intelligent machine and human being. This paper deals with the mechanical aspects of "Face Robot" that produces facial expressions in order to express the artificial emotions as in human being. We select the flexible microactuator (FMA) driven by air pressure for the sake of moving the control points of face robot corresponding to action units (AUs), and then we design and construct a 3 dimensional human-face-like robot. Then we undertake recognition test by showing the face images of 6 basic facial expressions expressed on the face robot, and the results show us that the correct recognition ratio accomplishes 83.3% for 6 basic facial expressions. This high ratio indicates the possibility of the face robot as a "KANSEI" communication media between intelligent machine and human being. >

Recognition of facial expressions using conceptual fuzzy sets

Abstract: A facial expression is a vague concept that is difficult to describe explicitly. Conceptual fuzzy sets (CFSs), which have the ability to explicitly represent vague concepts, are realized using bidirectional associative memories, and a multilayer structured CFS represents the meaning of a concept by various expressions in each layer. Multilayered reasoning in CFS has the capability of simultaneous abstract and concrete representation and of simultaneous top-down and bottom-up processing. CFS has been applied to the recognition of facial expressions and shown to achieve context-sensitive recognition. >

Biologically Motivated Approach to Face Recognition

Abstract: A biologically motivated compute intensive approach to computer vision is developed and applied to the problem of face recognition. The approach is based on the use of two-dimensional Gabor functions that fit the receptive fields of simple cells in the primary visual cortex of mammals. A descriptor set that is robust against translations is extracted by a global reduction operation and used for a search in an image database. The method was applied on a database of 205 face images of 30 persons and a recognition rate of 94% was achieved.

A basic study of dynamic recognition of human facial expressions

Abstract: In order to develop an active human interface (AHI) for heart-to-heart communication between machine and human being, we've been investigating the machine recognition of human emotions from facial expressions. This paper aims at clarifying the difference between static and dynamic recognition of facial expressions and investigating the characteristics of their dynamic recognition. In the first place, we obtain the 11 facial images for each of 6 basic expressions, sequentially changing from neutral to one of basic expressions (apex). Then, for comparison, we undertake static visual recognition tests by showing facial images to subjects, and then, by using the sequentially changing images, we also perform dynamic recognition tests. The comparison between static and dynamic recognition results reveals no difference. We further prepare two types of sequential facial images; time-wise sequential images changing from neutral to apex (UP) and from apex to neutral (DOWN). The results of dynamic recognition tests by using UP and DOWN sequential images reveal the fact that the recognition point in changing facial expression differs between the UP and DOWN time-sequential face image data, i.e., there exists a hysteresis in human dynamic recognition of facial expressions. >

Illumination-Invariant Face Recognition with a Contrast Sensitive Silicon Retina

Abstract: Changes in lighting conditions strongly effect the performance and reliability of computer vision systems. We report face recognition results under drastically changing lighting conditions for a computer vision system which concurrently uses a contrast sensitive silicon retina and a conventional, gain controlled CCO camera. For both input devices the face recognition system employs an elastic matching algorithm with wavelet based features to classify unknown faces. To assess the effect of analog on-chip preprocessing by the silicon retina the CCO images have been "digitally preprocessed" with a bandpass filter to adjust the power spectrum. The silicon retina with its ability to adjust sensitivity increases the recognition rate up to 50 percent. These comparative experiments demonstrate that preprocessing with an analog VLSI silicon retina generates image data enriched with object-constant features.

Fast face recognition method using high order autocorrelations

Abstract: We describe both the implementation and the evaluation of a face recognition method The feature extraction algorithm is based on the computation of local autocorrelation coefficients The main characteristics of these coefficients are simplicity of computation and a built-in translational invariance, which allows the system to respond in real time The classification is realized by conventional methods; namely least square discriminant mapping and linear discriminant analysis, which may be implemented with hardware neural networks We tested the system on a database of 11600 images of 116 persons The simulations show peak recognition rates of up to 98% and a satisfactory rejection vs recognition ratio

Knowledge-based face recognition using neural networks

Abstract: In this paper, we propose a method for facial feature extraction and recognition algorithm based on neural networks. First we separate the face part from the captured image based on the fact that the face image is located in the center of an input image and the background is relatively uniform. Then we obtain 4 normalized features from the extracted face image. For face recognition, we use the backpropagation technique of the neural networks. The proposed knowledge-based technique recognizes 14 persons correctly.

Analysis and synthesis of facial expression using high-definition wire frame model

Abstract: We propose the method to extract facial expression parameter with quantitative analysis of 9-dimensional movement of facial surface that arises from real human expression, using high-definition model. By dividing the face surface to some regions having similar characteristics of movement and modeling with 3-dimensional round surface in each region respectively, we can derive the muscle control parameters and the rule of movement. Considering the findings of this analysis we propose also a synthesis method of real facial expression image based on this high-definition model. >

A computational model of artificial emotion by using harmony theory and genetic algorithm

Abstract: This paper deals with a computational model of artificial emotion for "Active Human Interface" that generates emotion and facial expressions from the emotional evaluation state of external stimuli given to the model using the harmony theory, neural network and genetic algorithm. The harmony theory, a type of Boltzmann machine, is employed in this paper, and for this network system, we show a method of learning six basic emotions (joy, anger, sadness, fear, disgust and surprise). We also formulate schemata connecting emotional evaluation states and facial expressions consisting three facial components (eye, eyebrow and mouth). Simulation results show the successful emotion generation demonstrating the effectiveness of the genetic algorithm learning. >

What kinds of facial features are used in face retrieval

Abstract: An experimental evaluation of a facial image retrieval system showed that our previously proposed "context-driven retrieval mechanism" facilitated retrieval (or externalization) of ambiguous as well as better defined facial images. Some features, such as face shape, eyebrow tilt, and eye shape, were found to be more salient than others. The context-driven retrieval mechanism, while maintaining the relative importance of facial features, facilitated retrieval by reducing the variance in the less salient features. >

Synthetic face expressions generated by self organizing maps

Abstract: We describe a system for reproducing and generating face expressions by representing the significant segments on low-dimensional topology conserving maps. Our aim is to develop a representation of face movements which is derived exclusively from image data and contains no physiological information. The target of this research is real-time animation of face images which can be applied in advanced man-machine interfaces or face recognition tasks. There are several hard problems associated with this project: First, the significant segments of face expressions have to be be identified and a representation has to be found which allows to play back a sequence of facial expressions. Second, the system has to be trainable for the dependencies among the significant segments in order to develop an internal model for effects like variable lighting or small changes in orientation. We show that topology conserving maps and their extension to local linear mapping (LLM) networks allow a very efficient representation of critical face segments and that they are able to produce a sequence of realistic expressions in real time.

Shape classification: towards a mathematical description of the face

Abstract: Recent advances in biostereometric techniques have led to the quick and easy acquisition of 3D data for facial and other biological surfaces. This has led facial surgeons to express dissatisfaction with landmark-based methods for analysing the shape of the face which use only a small part of the data available, and to seek a method for analysing the face which maximizes the use of this extensive data set. Scientists working in the field of computer vision have developed a variety of methods for the analysis and description of 2D and 3D shape. These methods are reviewed and an approach, based on differential geometry, is selected for the description of facial shape. For each data point, the Gaussian and mean curvatures of the surface are calculated. The performance of three algorithms for computing these curvatures are evaluated for mathematically generated standard 3D objects and for 3D data obtained from an optical surface scanner. Using the signs of these curvatures, the face is classified into eight 'fundamental surface types' - each of which has an intuitive perceptual meaning. The robustness of the resulting surface type description to errors in the data is determined together with its repeatability. Three methods for comparing two surface type descriptions are presented and illustrated for average male and average female faces. Thus a quantitative description of facial change, or differences between individual's faces, is achieved. The possible application of artificial intelligence techniques to automate this comparison is discussed. The sensitivity of the description to global and local changes to the data, made by mathematical functions, is investigated. Examples are given of the application of this method for describing facial changes made by facial reconstructive surgery and implications for defining a basis for facial aesthetics using shape are discussed. It is also applied to investigate the role played by the shape of the surface in facial recognition.

Neural network recognition of human face images stored in the database

Abstract: The author describes an investigation of human face recognition using neural networks. The investigation is the basis for retrieval and management of human face images stored in the database. The database is similar to FBI or many other law enforcement agencies databases. The goal of the investigation is twofold. First, assuming that an existing database of the front face and profile images is available, the requirement is to know whether the neural network trained on the front image and the profile can recognize any other images of the same person. Second, a minimal set of snapshots of each person is desired consisting of at least the front face and profile, which are needed to train a neural network so that the trained network can then recognize many other snapshots of the same person. The research prototype of a postrelational database management system (DBMs) is discussed called CHINOOK being implemented at the University of Colorado at Colorado Springs. CHINOOK is intended to manage databases of digitized images and digitized one-dimensional data as well as text and tables. CHINOOK is intended to support the retrieval of images by their content. >

A facial feature communications interface for the non-vocal

Abstract: A viable communications interface for the nonvocal handicapped that is a by-product of recent face recognition research is described. The system uses either the Karhunen-Loeve transform (KLT) or the discrete cosine transform (DCT) to encode video camera images on a subject's face. The encoding differentiates among three expressions: tongue out which is classified as 'yes'; mouth open, which is classified as 'no'; and mouth closed, which is classified as 'null'. A review of the theory of the KLT and DCT is included. >

Recognition of and reasoning about facial expressions using fuzzy logic

Abstract: In the study of the linguistic modeling of facial images we have been previously concerned with deriving qualitative descriptions such as "big eyes, long hair" of face components. To enhance this system we extend our approach at deriving higher level, qualitative descriptions. In particular, we focus on describing facial expressions. Our approach is that of qualitative modeling based on fuzzy number modeling. The result of this modeling method is a collection of fuzzy if-then rules obtained from input-output data. The input data consists of measurements of the movement of facial parts associated to different facial expressions. The output data consists of scores for face images collected using a questionnaire. In this paper, we show the modeling result obtained from this method for the facial expression "happy". While the modeling results are satisfactory the initial recognition results are limited, due in part to the absence of the models for the remaining facial expressions. >

Novel approach to human face recognition

Abstract: This paper presents a new method of human face recognition based on a novel algebraic feature extraction method. An input human face image is First transformed into a standard image; Then, the projective feature vectors of the standard image are extracted by projecting it onto the optimal discriminant projection vectors; Finally, face image recognition is completed by classifying these projective feature vectors. Experimental results showed that the present method is effective.