Hybrid Neural Networks for Pattern Recognition
Reads0
Chats0
TLDR
The enhanced hybrid network is proposed that organizes the middle layer effectively by using the enhanced ART1 network adjusting the vigilance parameter dynamically according to the similarity between patterns to improve the recognition success rate.Abstract:
The hybrid neural networks have characteristics such as fast learning times, generality, and simplicity, and are mainly used to classify learning data and to model non-linear systems. The middle layer of a hybrid neural network clusters the learning vectors by grouping homogenous vectors in the same cluster. In the clustering procedure, the homogeneity between learning vectors is represented as the distance between the vectors. Therefore, if the distances between a learning vector and all vectors in a cluster are smaller than a given constant radius, the learning vector is added to the cluster. However, the usage of a constant radius in clustering is the primary source of errors and therefore decreases the recognition success rate. To improve the recognition success rate, we proposed the enhanced hybrid network that organizes the middle layer effectively by using the enhanced ART1 network adjusting the vigilance parameter dynamically according to the similarity between patterns. The results of experiments on a large number of calling card images showed that the proposed algorithm greatly improves the character extraction and recognition compared with conventional recognition algorithms.read more
Citations
More filters
Posted Content
Deep Neural Networks for Pattern Recognition.
TL;DR: The basic structure of deep neural networks that simulate human neural networks are introduced and the operational processes and applications of conditional generative adversarial networks are identified, which are being actively researched based on the bottom-up and top-down mechanisms of the human visual perception process.
Proceedings ArticleDOI
Neural network target identification system for false alarm reduction
TL;DR: The test results show that the ATR system was successful in substantially reducing the false positive rate when tested on a sonar image dataset and the parameter optimizations process which adapts the system to various targets and datasets are discussed.
Proceedings ArticleDOI
Deep neural network for precision multi-band infrared image segmentation
TL;DR: The implementation of a Deep Neural Network (DNN) to perform precision segmentation of multi-band IR video images is presented and single pixel accuracy has been achieved in IR target boundary segmentation using DNNs.
Journal ArticleDOI
Improving data quality using a deep learning network
TL;DR: When applying LSTM to data quality improvement in IoT environment where data are collected simultaneously from several sensors, it is suggested that it is effective to construct L STM individually for each sensor accuracy.
Proceedings ArticleDOI
Intelligent multi-spectral IR image segmentation
Thomas Lu,Andrew Luong,Stephen Heim,Maharshi Patel,Kang (Frank) Chen,Tien-Hsin Chao,Edward Chow,Gilbert Torres +7 more
TL;DR: The neural network based segmentation method compared to the Wavelet-threshold and Grab-Cut methods has shown increased accuracy and robustness of this segmentation scheme for multi-spectral IR images.
References
More filters
Book
Fuzzy Set Theory - and Its Applications
TL;DR: The book updates the research agenda with chapters on possibility theory, fuzzy logic and approximate reasoning, expert systems, fuzzy control, fuzzy data analysis, decision making and fuzzy set models in operations research.
Journal ArticleDOI
Fuzzy Set Theory and Its Applications
TL;DR: In this paper, a new book about fuzzy set theory and its applications is presented, which can be used to explore the knowledge of the knowledge in a new way, even for only few minutes to read a book.
Journal ArticleDOI
Universal approximation using radial-basis-function networks
Jooyoung Park,Irwin W. Sandberg +1 more
TL;DR: It is proved thatRBF networks having one hidden layer are capable of universal approximation, and a certain class of RBF networks with the same smoothing factor in each kernel node is broad enough for universal approximation.
Proceedings ArticleDOI
Effects of moving the centers in an RBF network
TL;DR: Borders for the gradient and Hessian of the error considered as a function of centers for networks of fixed size are calculated and it is possible to know by how much one can reduce the error by changing the centers.
Journal ArticleDOI
An Efficient Extraction of Pulmonary Parenchyma in CT Images using Connected Component Labeling
TL;DR: The proposed method is based on the calculation of the connected component and the centroid of the image which helps to extract the lungs part from the image retaining all the parts of the original lungs image.