Catálogo de publicaciones - libros

In video-based of face recognition applications, the What-and-Where Fusion Neural Network (WWFNN) has been shown to reduce the generalization error by accumulating a classifier’s predictions over time, according to each individual in the environment. In this paper, three ARTMAP variants – fuzzy ARTMAP, ART-EMAP (Stage 1) and ARTMAP-IC – are compared for the classification of faces detected in the WWFNN. ART-EMAP (stage 1) and ARTMAP-IC expand on the well-known fuzzy ARTMAP by using distributed activation of category neurons, and by biasing distributed predictions according to the number of time these neurons are activated by training set patterns. The average performance of the WWFNNs with each ARTMAP network is compared to the WWFNN with a reference -NN classifier in terms of generalization error, convergence time and compression, using a data set of real-world video sequences. Simulations results indicate that when ARTMAP-IC is used inside the WWFNN, it can achieve a generalization error that is significantly higher (about 20% on average) than if fuzzy ARTMAP or ART-EMAP is used. Indeed, ARTMAP-IC is less discriminant than the two other ARTMAP networks in cases with complex decision bounderies, when the training data is limited and unbalanced, as found in complex video data. However, ARTMAP-IC can outperform the others when classes are designed with a larger number of training patterns.

This paper proposes a new method for face recognition based on a multiresolution analysis tool called Digital Curvelet Transform. Multiresolution ideas notably the wavelet transform have been profusely employed for addressing the problem of face recognition. However, theoretical studies indicate, digital curvelet transform to be an even better method than wavelets. In this paper, the feature extraction has been done by taking the curvelet transforms of each of the original image and its quantized 4 bit and 2 bit representations. The curvelet coefficients thus obtained act as the feature set for classification. These three sets of coefficients from the three different versions of images are then used to train three Support Vector Machines. During testing, the results of the three SVMs are fused to determine the final classification. The experiments were carried out on three well known databases, viz., the Georgia Tech Face Database, AT&T “The Database of Faces” and the Essex Grimace Face Database.

A novel method for face recognition based on some biological aspects of infant vision is proposed in this paper. The biological hypotheses of this method are based on the role of the response to low frequencies at early stages, and some conjectures concerning how an infant detects subtle features (stimulating points) from a face. In order to recognize a face from different images of it we make use of a bank of dynamic associative memories (DAM). As the infant vision responds to low frequencies of the signal, a low-filter is first used to remove high frequency components from the image. We then detect subtle features in the image by means of a random feature selection detector. At last, the network of DAMs is fed with this information for training and recognition. To test the accuracy of the proposal a benchmark of faces is used.

In this paper, we propose a novel approach for facial expression analysis and recognition. The proposed approach relies on the distance vectors retrieved from 3D distribution of facial feature points to classify universal facial expressions. Neural network architecture is employed as a classifier to recognize the facial expressions from a distance vector obtained from 3D facial feature locations. Facial expressions such as anger, sadness, surprise, joy, disgust, fear and neutral are successfully recognized with an average recognition rate of 91.3%. The highest recognition rate reaches to 98.3% in the recognition of surprise.

This paper presents a framework that combines a robust facial features location with an elliptical face modelling to measure user’s intention and point of attention. The most important facial feature points are automatically detected using a statistically anthropometric face model. After observing the structural symmetry of the human face and performing some anthropometric measurements, the system is able to build a model that can be used in isolating the most important facial feature areas: mouth, eyes and eyebrows. Combination of different image processing techniques are applied within the selected regions for detecting the most important facial feature points.

A model based approach is used to estimate the 3D orientation of the human face. The shape of the face is modelled as an ellipse assuming that the human face aspect ratio (ratio of the major to minor axes of the 3D face ellipse) is known. The elliptical fitting of the face at the image level is constrained by the location of the eyes which considerable increase the performance of the system.

The system is fully automatic and classifies rotation in all-view direction, detects eye blinking and eye closure and recovers the principal facial features points over a wide range of human head rotations. Experimental results using real images sequences demonstrates the accuracy and robustness of the proposed solution.

This paper presents an iris recognition technique based on the zigzag collarette region for segmentation and asymmetrical support vector machine to classify the iris pattern. The deterministic feature sequence extracted from the iris images using the 1D log-Gabor filters is applied to train the support vector machine (SVM). We use the multi-objective genetic algorithm (MOGA) to optimize the features and also to increase the overall recognition accuracy based on the matching performance of the tuned SVM. The traditional SVM is modified to an asymmetrical SVM to treat the cases of the False Accept and the False Reject differently and also to handle the unbalanced data of a specific class with respect to the other classes. The proposed technique is computationally effective with recognition rates of 97.70 % and 95.60% on the ICE (Iris Challenge Evaluation) and the WVU (West Virginia University) iris datasets respectively.

Automated brain MR image segmentation is a challenging pattern recognition problem that received significant attention lately. The most popular solutions involve fuzzy c-means (FCM) or similar clustering mechanisms. Several improvements have been made to the standard FCM algorithm, in order to reduce its sensitivity to Gaussian, impulse, and intensity non-uniformity noises. This paper presents a modified FCM-based method that targets accurate and fast segmentation in case of mixed noises. The proposed method extracts a scalar feature value from the neighborhood of each pixel, using a context dependent filtering technique that deals with both spatial and gray level distances. These features are clustered afterwards by the histogram-based approach of the enhanced FCM algorithm. Results were evaluated based on synthetic phantoms and real MR images. Test experiments revealed that the proposed method provides better results compared to other reported FCM-based techniques. The achieved segmentation and the obtained fuzzy membership values represent excellent support for deformable contour model based cortical surface reconstruction methods.

The aim of this paper is to evaluate and compare the performance of three convergence index () filters when applied to the enhancement of chest radiographs, aiming at the detection of lung nodules. One of these filters, the sliding band filter (), is the proposal of an innovative operator, while the other two class members, the iris filter () and the adaptive ring filter (), are already known in this application. To demonstrate the adequacy of the new filter for the enhancement of chest x-ray images, we calculated several figures of merit with the goal of comparing (i) the contrast enhancement capability of the filters, and (ii) the behavior of the filters for the detection of lung nodules. The results obtained for 154 images with nodules of the JSRT database show that the outperforms both the and . The proposed filter demonstrated to be a promising enhancement method, thus justifying its use in the first stage of a computer-aided diagnosis system for the detection of lung nodules.

This paper presents a new method to temporally track the principal coronary arteries in an X-ray fluoroscopy setting. First, the principal coronary artery centerlines are extracted at a first time instant. Secondly, in order to estimate the centerline coordinates in subsequent time frames, a pyramidal Lucas-Kanade optical flow approach is used. Finally, an active contour model coupled with a gradient vector flow (GVF) formulation is used to deform the estimated centerline coordinates towards the actual medial axis positions. The algorithm’s effectiveness has been evaluated on 38 monoplane images from three patients and the temporal tracking lasted over one cardiac cycle. The results show that the centerlines were correctly tracked in 92% of the image frames.

Female breast cancer is a major cause of death in western countries. Several computer techniques have been developed to aid radiologists to improve their performance in the detection and diagnosis of breast abnormalities. In Point Pattern Analysis, there is a statistic known as Ripley’s function that is frequently applied to Spatial Analysis in Ecology, like mapping specimens of plants. This paper proposes a new way in applying Ripley’s function in order to distinguish Mass and Non-Mass tissues from mammogram images. The features of each image are obtained through the calculate of that function. Then, the samples gotten are classified through a Support Vector Machine (SVM) as Mass or Non-Mass tissues. SVM is a machine-learning method, based on the principle of structural risk minimization, which performs well when applied to data outside the training set. Another way of computing Ripley’s function, using concentric rings instead of a circle, is also examined. The best result achieved was 94.25% of accuracy, 94.59% of sensitvity and 94.00% of specificity.