Proceedings of the 12th IEEE Workshop on Neural Networks for Signal Processing
Cơ quản chủ quản: N/A
Các bài báo tiêu biểu
Signal reconstruction from sampled data using neural network
- Trang 707-715
For reconstructing a signal from sampling data, the method based on Shannon's sampling theorem is usually employed. The reconstruction error appears when the signal does not satisfy the Nyquist condition. This paper proposes a new reconstruction method by using a linear perceptron and multilayer perceptron as FIR filter. The perceptron, which has weights obtained by learning when adapting the orig......
#Signal reconstruction #Neural networks #Image reconstruction #Sampling methods #Finite impulse response filter #Frequency #Adaptive filters #Image sampling #Information retrieval #Physics
Feature selection for off-line recognition of different size signatures
- Trang 355-364
The aim of this work is to select a set of features, which have good performance to solve the problem of signature recognition of different sizes. The signature database was formed for three sizes of signatures per user, small, median and big. This study uses structural features, pseudo-dynamic features and five moment groups. The feature selection method chosen is the one that select the best ind......
#Spatial databases #Handwriting recognition #Feature extraction #Data analysis #Data mining #Cameras #Forgery
Accurate estimation of the signal baseline in DNA chromatograms
- Trang 35-44
Estimating accurately the varying baseline level in different parts of a DNA chromatogram is a challenging and important problem for accurate base-calling. We are formulating the problem in a statistical learning framework and propose an Expectation-Maximization algorithm for its solution. In addition we also present a faster, iterative histogram based method for estimating the background of the s......
#Statistical learning #Bayesian methods #Signal resolution #Signal to noise ratio #Data mining #Digital signal processing #DNA computing #Expectation-maximization algorithms #Iterative methods #Histograms
On learning feedforward neural networks with noise injection into inputs
- Trang 149-158
Injecting noise to the inputs during the training of feedforward neural networks (FNN) can improve their generalization performance remarkably. Reported works justify this fact arguing that noise injection is equivalent to a smoothing regularization with the input noise variance playing the role of the regularization parameter. The success of this approach depends on the appropriate choice of the ......
#Neural networks #Feedforward neural networks #Cost function #Neurons #Function approximation #Smoothing methods #Pattern classification #Approximation algorithms #Electronic mail #Fuzzy control
Improvements on continuous unsupervised sleep staging
- Trang 687-695
We report improvements on automatic continuous sleep staging using hidden Markov models (HMM). Contrary to our previous efforts, we trained the HMMs on data from single sleep labs instead of generalizing to data from diverse sleep labs. Our totally unsupervised approach detects the cornerstones of human sleep (wakefulness, deep and rem sleep) with around 80% accuracy based on data from a single EE......
#Sleep #Hidden Markov models #Electroencephalography #Electromyography #Humans #Electrooculography #Artificial intelligence #Brain modeling #Electrodes #Reflection
Blind source separation with different sensor spacing and filter length for each frequency range
- Trang 465-474
This paper presents a method for blind source separation using several separating subsystems whose sensor spacing and filter length can be configured individually. Each subsystem is responsible for source separation of an allocated frequency range. With this mechanism, we can use appropriate sensor spacing as well as filter length for each frequency range. We obtained better separation performance......
#Blind source separation #Frequency #Source separation #Finite impulse response filter #Independent component analysis #Radio spectrum management #Reverberation #Laboratories #Sensor systems #Jamming
Dynamic Bayesian network based speech recognition with pitch and energy as auxiliary variables
- Trang 637-646
Pitch and energy are two fundamental features describing speech, having importance in human speech recognition. However, when incorporated as features in automatic speech recognition (ASR), they usually result in a significant degradation on recognition performance due to the noise inherent in estimating or modeling them. We show experimentally how this can be corrected by either conditioning the ......
#Bayesian methods #Speech recognition #Automatic speech recognition #Degradation #Hidden Markov models #Acoustic emission #Artificial intelligence #Humans #Speech enhancement #Network topology
Towards the introduction of human perception in a natural scene classification system
- Trang 385-394
We develop a method to optimize a machine-based semantic categorization of natural images according to human perception. First, the categories are determined through a psychophysical experiment. The similarity matrices obtained from the human responses are analyzed by a multidimensional scaling technique called curvilinear component analysis (CCA). The same is done with an automatic image indexing......
#Humans #Layout #Image databases #Gabor filters #Indexing #Image retrieval #Psychology #Multidimensional systems #Multimedia databases #Navigation
Language model adaptation in speech recognition using document maps
- Trang 627-636
We present speech experiments that were carried out to evaluate a topically focusing language model in large vocabulary speech recognition. An ordered topical clustering is first computed as a self-organized mapping of a large document collection. Language models are then trained for each text cluster or for several neighboring clusters. The obtained organized collection of language models is effi......
#Natural languages #Adaptation model #Speech recognition #Vocabulary #Probability #Intelligent networks #Neural networks #Speech analysis #Databases #Ultraviolet sources
Modified Kalman filter based method for training state-recurrent multilayer perceptrons
- Trang 219-228
Kalman filter based training algorithms for recurrent neural networks provide a clever alternative to the standard backpropagation in time. However, these algorithms do not take into account the optimization of the hidden state variables of the recurrent network. In addition, their formulation requires Jacobian evaluations over the entire network, adding to their computational complexity. We propo......
#Multilayer perceptrons #Backpropagation algorithms #Recurrent neural networks #Signal processing algorithms #Computational complexity #Jacobian matrices #Kalman filters #Neural networks #Convergence #Neural engineering