International Journal of Pattern Recognition and Artificial Intelligence

Typical testors are useful tools for feature selection and for determining feature relevance in supervised classification problems. Nowadays, computing all typical testors of a training matrix is very expensive; all reported algorithms have exponential complexity depending on the number of columns in the matrix. In this paper, we introduce the faster algorithm BR (Boolean Recursive), called fast-BR algorithm, that is based on elimination of gaps and reduction of columns. Fast-BR algorithm is designed to generate all typical testors from a training matrix, requiring a reduced number of operations. Experimental results using this fast implementation and the comparison with other state-of-the-art related algorithms that generate typical testors are presented.

Recognition of natural gestures is a key issue in many applications including videogames and other immersive applications. Whatever is the motion capture device, the key problem is to recognize a motion that could be performed by a range of different users, at an interactive frame rate. Hidden Markov Models (HMM) that are commonly used to recognize the performance of a user however rely on a motion representation that strongly affects the overall recognition rate of the system. In this paper, we propose to use a compact motion representation based on Morphology-Independent features and we evaluate its performance compared to classical representations. When dealing with 15 very similar upper limb motions, HMM based on Morphology-Independent features yield significantly higher recognition rate (84.9%) than classical Cartesian or angular data (70.4% and 55.0%, respectively). Moreover, when the unknown motions are performed by a large number of users who have never contributed to the learning process, the recognition rate of Morphology-Independent input feature only decreases slightly (down to 68.2% for a HMM trained with the motions of only one subject) compared to other features (25.3% for Cartesian features and 17.8% for angular features in the same conditions). The method is illustrated through an interactive demo in which three virtual humans have to interactively recognize and replay the performance of the user. Each virtual human is associated with a HMM recognizer based on the three different input features.

License plate recognition (LPR) is one of the most important elements affecting intelligent transportation systems. A number of LPR techniques have been proposed. Humans are good target recognition systems. In other words, humans easily recognize common objects. In this paper, the researchers present a novel method of recognizing Chinese license plates. The method is based on the Human Vision Attention Mechanism (HVAM) and uses Chinese license plates as the targets. The research consists of three stages. The first stage involved finding and identifying license plates in videos of moving vehicles. The second stage separated each license plate into the seven characters. In the third stage, the character recognizer extracted some salient features of Chinese characters and used a multi-stage classifier to recognize each character on the license plate. In the experiment locating license plates, 1176 images taken from various scenes and conditions were employed. The method failed to identify the license plates in only 27 of the images; resulting in a license plate location rate of success of 97.7%. In the experiment for identifying license characters, 1149 images were used, from which license plates had been successfully located. The method failed to identify the characters in 45 of these images giving a success rate of 96.1%. Combining the above two rates, the overall rate of success for our LPR is 93.9%.

The motor vehicle is an important way of transportation for modern people, and its license plate is just like our identification cards which can be used for effective management of motor vehicles. Hence, the development of a recognition system for license plates can reduce the workload of managing motor vehicles. A license plate recognition system based on computer vision often causes recognition errors due to the plate's angle problem and thus needs to be assisted by manual recognition. In this study, a recovery method for license plate images with large angles is proposed based on perspective projection to increase the recognition accuracy. The experimental results show that applying the angle recovery method to a license plate recognition system can reduce its errors, especially for license plate images with large angles. For the case of small angles, a recognition rate of 89% can be achieved by the angle recovery method, slightly higher than that of the plane rotation method (86%). For the case of a wide range of different angles, the average recognition rate achieved by the angle recovery method is 87.5%, much higher than that of the plane rotation method (57.5%). Thus, the angle recovery method is effective for enhancing the accuracy of license plate recognition systems.

To improve the accuracy of visual speech recognition systems, selection of visual features is of fundamental importance. Prominent features, which are of maximum relevance for speech classification, need to be selected from a large set of extracted visual attributes. Existing methods apply feature reduction and selection techniques on image pixels constituting region-of-interest (ROI) to reduce data dimensionality. We propose application of feature selection methods on geometrical features to select the most dominant physical features. Two techniques, Minimum Redundancy Maximum Relevance (mRMR) and Correlation-based Feature Selection (CFS), have been applied on the extracted visual features. Experimental results show that recognition accuracy is not compromised when a few selected features from the complete visual feature set are used for classification, thereby reducing processing time and storage overheads considerably. Results are compared with performance of principal components obtained by application of Principal Component Analysis (PCA) on our dataset. Our set of selected features outperforms the PCA transformed data. Results show that the center and corner segments of the mouth are major contributors to visual speech recognition. Teeth pixels are shown to be a prominent visual cue. It is also seen that lip width contributes more towards visual speech recognition accuracy as compared to lip height.

A new approach for ensemble construction based on restricting a set of weights of examples in training data to avoid overfitting is proposed in the paper. The algorithm called EPIBoost (Extreme Points Imprecise Boost) applies imprecise statistical models to restrict the set of weights. The updating of the weights within the restricted set is carried out by using its extreme points. The approach allows us to construct various algorithms by applying different imprecise statistical models for producing the restricted set. It is shown by various numerical experiments with real data sets that the EPIBoost algorithm may outperform the standard AdaBoost for some parameters of imprecise statistical models.

The number of inner points excluded in an initial convex hull (ICH) is vital to the efficiency getting the convex hull (CH) in a planar point set. The maximum inscribed circle method proposed recently is effective to remove inner points in ICH. However, limited by density distribution of a planar point set, it does not always work well. Although the affine transformation method can be used, it is still hard to have a better performance. Furthermore, the algorithm mentioned above fails to deal with the exceptional distribution: the gravity centroid (GC) of a planar point set is outside or on the edge formed by the extreme points in ICH. This paper considers how to remove more inner points in ICH when GC is inside of ICH and completely process the case which mentioned above. Further, we presented a complete algorithm architecture: (1) using the ellipse and elasticity ellipse methods (EM and EEM) to remove more inner points in ICH and process the cases: GC is inside or outside of ICH. (2) Using the traditional methods to process the situation: the initial centroid is on the edge in ICH. It is adaptive to more data sets than other algorithms. The experiments under seven distributions show that the proposed method performs better than other traditional algorithms in saving time and space.

The usefulness of an image database depends on whether the image of interest can be easily located. Feature extraction is a crucial step of image retrieval. The well-known SIFT descriptor is a keypoint-based image feature. It can be used to robustly find the same objects in different images, i.e. achieve the object recognition task. However, it is not effective on the image retrieval task, i.e. finding images with similar content. To improve the SIFT algorithm, we propose a robust image retrieval algorithm based on the integration of keypoints and edges information. Our approach is robust to translation, rotation and partial occlusion of the object. Experimental results indicate that the proposed algorithm is effective and outperforms the SIFT algorithm.

In this paper, a genetic algorithm (GA)-based approach to estimate the fundamental matrix is presented. The aim of the proposed GA-based algorithm is to reduce the effect of noise and outliers in the corresponding points which affect the accuracy of the estimated fundamental matrix. Although in the proposed approach the GA is allowed to select the significant among all detected points, on the average half of the matched points have been determined to give optimum estimation of the fundamental matrix. Experiments with synthetic and real data show that the proposed approach is accurate especially in the presence of a high percentage of outliers. The proposed GA can always obtain good results in both high and low detailed images. Even for low detailed images which have a small number of matched points available to estimate the fundamental matrix, the proposed GA outperformed other methods.

As is known, kernel methods are developed to handle nonlinear classification. However, we have found that based on vector representation of face images, it is not easy to improve the performance of LDA-based methods by incorporating the kernel trick. Actually, in the case that the dimensionality of face images in vector form is far greater than the number of samples, linear classifiers are adequate and it is illogical to increase dimensions of vectors using kernel methods. In order to give full play to the strong point of the kernel technique for manipulating the nonlinearity of pattern distribution, we propose a new image feature extraction method for face recognition, called kernel-based two-dimensional Fisher discriminant analysis (K2DFDA), which deals with a face image directly as a matrix, instead of a stacked vector from rows or columns of the image. Moreover, we present a kernel parameter optimization scheme for K2DFDA, based on the maximum margin criterion and the damped Newton's method. Experimental results show the effectiveness of K2DFDA and its parameter optimization scheme.