In this paper, we develop an intelligent application based neuralnetworks and imageprocessing to recognize license plate for car management. Through the license recognition, the car number composed of English alphab...
详细信息
ISBN:
(纸本)9783642166952
In this paper, we develop an intelligent application based neuralnetworks and imageprocessing to recognize license plate for car management. Through the license recognition, the car number composed of English alphabets and digitals is readable for computers. Recognition of license is processed in two stages including feature extraction and recognition. The feature extraction contains the image locating, segmentation of the region of interest (ROT). Then the extracted ROIs are fed to a trained neural network for recognition. The neural network is a three-layer feed-forward neural network. Test images are produced from real parking lots. There are 500 images of car plates with tile, zooming and various lighting conditions, for verification. The experiment results show that the ratio of successful locating of license plate is around 96.8%, and the ratio of successful segmentation is 91.1%. The overall successful recognition ratio is 87.5%. Therefore, the experimental result shows that the proposed method works effectively, and simultaneously to improve the accuracy for the recognition. This system improves the performance of automatic license plate recognition for future ITS applications.
Visual pattern recognition is a complex problem, and it has proven difficult to achieve satisfactorily in standard three-layer feed-forward artificialneuralnetworks. For this reason, an increasing number of research...
详细信息
ISBN:
(纸本)9789896740283
Visual pattern recognition is a complex problem, and it has proven difficult to achieve satisfactorily in standard three-layer feed-forward artificialneuralnetworks. For this reason, an increasing number of researchers are using networks whose architecture resembles the human visual system. These biologically based networks are bidirectionally connected, use receptive fields, and have a hierarchical structure, with the input layer being the largest layer, and consecutive layers getting increasingly smaller. These networks are large and complex, and therefore run a risk of getting overfitted during learning, especially if small training sets are used, and if the input patterns are noisy. Many data sets, such as, for example, handwritten characters, are intrinsically noisy. The problem of overfitting is aggravated by the tendency of error-driven learning in large networks to treat all variations in the noisy input as significant. However, there is one way to balance off this tendency to overfit, and that is to use a mixture of learning algorithms. In this study, we ran systematic tests on handwritten character recognition, where we compared generalization performance using a mixture of Hebbian learning and error-driven learning with generalization performance using pure error-driven learning. Our results indicate that injecting even a small amount of Hebbian learning, 0.01 %, significantly improves the generalization performance of the network.
Hierarchical Temporal Memory (HTM) is an emerging computational paradigm consisting of a hierarchically connected network of nodes. The hierarchy models a key design principle of neocortical organization. Nodes throug...
详细信息
ISBN:
(纸本)9783642158216
Hierarchical Temporal Memory (HTM) is an emerging computational paradigm consisting of a hierarchically connected network of nodes. The hierarchy models a key design principle of neocortical organization. Nodes throughout the hierarchy encode information by means of clustering spatial instances within their receptive fields according to temporal proximity. Literature shows HTMs' robust performance on traditional machine learning tasks such as image recognition. Problems involving multi-variable time series where instances unfold over time with no complete spatial representation at any point in time have proven trickier for HTMs. We have extended the traditional HTMs' principles by means of a top node that stores and aligns sequences of input patterns representing the spatio-temporal structure of instances to be learned. This extended HTM network improves performance with respect to traditional HTMs in machine learning tasks whose input instances unfold over time.
Nowadays, automatic defect detection in Breast images which obtains from mommogram is very important in many diagnostic and therapeutic applications. This paper introduces a Novel automatic breast abnormality detectio...
详细信息
This paper presents an intelligent system to perform an erratic driving diagnosis. The proposed approach takes into account the analysis of the signals that could be acquired from modern on-board diagnostic systems (O...
详细信息
A novel method for open-close eye states detection, based on complex wavelet transform (CWT) and complex-valued artificialneural network (CVANN) is proposed in this study. Firstly, color information of images is used...
详细信息
The proceedings contain 81 papers. The special focus in this conference is on artificialneuralnetworks, Computational Intelligence and Machine Learning. The topics include: information theory related learning;optimi...
ISBN:
(纸本)9782874190445
The proceedings contain 81 papers. The special focus in this conference is on artificialneuralnetworks, Computational Intelligence and Machine Learning. The topics include: information theory related learning;optimization of parametrized divergences in fuzzy c-means;statistical dependence measure for feature selection in microarray datasets;sparsity issues in self-organizing-maps for structures;multi-goal path planning using self-organizing map with navigation functions;negatively correlated echo state networks;reservoir regularization stabilizes learning of echo state networks with output feedback;a multi-kernel framework for inductive semi-supervised learning;training of multiple classifier systems utilizing partially labeled sequential data sets;recent trends in computational intelligence in life sciences;generalized functional relevance learning vector quantization;patch affinity propagation;multispectral image characterization by partial generalized covariance;anticipating rewards in continuous time and space with echo state networks and actor-critic design;application of stochastic recurrent reinforcement learning to index trading;a draughts player neural network that learns by reinforcement in a high performance environment;new conditioning model for robots;stability of neural network control for uncertain sampled-data systems;thresholds tuning of a neuro-symbolic net controlling a behavior-based robotic system;ensemble usage for more reliable policy identification in reinforcement learning;fisherman learning algorithm of the SOM realized in the CMOS technology;symbolic computing of LS-SVM based models;a post-processing strategy for SVM learning from unbalanced data;clustering data streams with weightless neuralnetworks;locating anomalies using Bayesian factorizations and masks;the importance of visualization in real-world machine learning applications;hierarchical clustering for graph visualization;a probabilistic approach to the visual exploration of G p
The proceedings contain 76 papers. The topics discussed include: emergent dynamics of information propagation in large networks;new applications and theoretical foundations of the dominance-based rough set approach;co...
ISBN:
(纸本)3642135285
The proceedings contain 76 papers. The topics discussed include: emergent dynamics of information propagation in large networks;new applications and theoretical foundations of the dominance-based rough set approach;consensus multiobjective differential crisp clustering for categorical data analysis;probabilistic rough entropy measures in image segmentation;distance based fast hierarchical clustering method for large datasets;vehicle classification based on soft computing algorithms;controlling computer by lip gestures employing neuralnetworks;computer animation system based on rough sets and fuzzy logic;adaptive phoneme alignment based on rough set theory;monitoring disease patients employing biometric sensors and rule-based data processing;content-based scene detection and analysis method for automatic classification of TV sports news;and combining multiple classification or regression models using genetic algorithms.
The proceedings contain 50 papers. The topics discussed include: moving targets: when data classes depend on subjective judgment, or they are crafted by an adversary to mislead pattern analysis algorithms - the cases ...
ISBN:
(纸本)3642143997
The proceedings contain 50 papers. The topics discussed include: moving targets: when data classes depend on subjective judgment, or they are crafted by an adversary to mislead pattern analysis algorithms - the cases of content based image retrieval and adversarial classification;bioinformatics contributions to data mining;bootstrap feature selection for ensemble classifiers;evaluating the quality of clustering algorithms using cluster path lengths;finding irregularly shaped clusters based on entropy;fuzzy conceptual clustering;mining concept similarities for heterogeneous ontologies;re-mining positive and negative association mining results;multi-agent based clustering: towards generic multi-agent data mining;describing data with the support vector shell in distributed environments;robust clustering using discriminant analysis;new approach in data stream association rule mining based on graph structure;fast training of neuralnetworks for image compression;and processing handwritten words by intelligent use of OCR results.
image based human pose recovery has many applications in different industries such as games, entertainment, physiological rehabilitation and biometrics. This paper presents a new pose estimation algorithm from monocul...
详细信息
image based human pose recovery has many applications in different industries such as games, entertainment, physiological rehabilitation and biometrics. This paper presents a new pose estimation algorithm from monocular images based on a nonlinear mapping of human silhouettes, coded using a collection of local image moments, to the pose space using a mixture of neuralnetworks (NN) regressors. All parameters are estimated automatically. Experiments and comparative results show a superior performance of the proposed method.
暂无评论