We consider optimization of hybrid imaging systems including a pupil mask for enhancing the depth of field and a digital deconvolution step. In a previous paper [Opt. Lett. 34, 2970 (2009)] we proposed an optimization criterion based on the signal-to-noise ratio of the restored image. We use this criterion in order to optimize different families of phase or amplitude masks and to compare them, ...

Neural networks are widely used as classifiers in many pattern recognition problems because of good generalization abilities, what is a crucial issue in any practical application. However, vast majority of neural network architectures demands a huge computational effort for the training process, what in turn limits such solutions from application in one important domain of pattern recognition, ...

We consider optimization of hybrid imaging systems including a phase mask for enhancing the depth of field and a digital deconvolution step. We propose an image quality criterion that takes into account the variability of the system's point-spread function along the expected defocus range and the noise enhancement induced by deconvolution. Considering the classical cubic phase mask as an exampl...

This paper is dedicated to the study of stochastic finite-time stability (SFTS) and control synthesis for a class of nonlinear Markovian jump stochastic systems with impulsive effects. By introducing a time-varying stochastic Lyapunov function with discontinuities at impulse times, an improved criterion for SFTS is derived in terms of linear matrix inequalities (LMIs). Based on the new SFTS cri...

The procedure of calculating Mel Frequency based Cepstral Coefficients (MFCC) is shown to resemble a three layer Multilayer Perceptron (MLP) like structure. Such an MLP is employed as a preprocessor in a hybrid HMM-MLP system, and the possibility of optimizing the whole system as a single entity, with respect to a suitable criterion, is pointed out. This system, together with the Maximum Mutual...

Approaches to machine intelligence based on brain models have stressed the use of neural networks for generalization. Here we propose the use of a hybrid neural network architecture that uses two kind of neural networks simultaneously: (i) a surface learning agent that quickly adapt to new modes of operation; and, (ii) a deep learning agent that is very accurate within a specific regime of oper...

In this paper we study a subset of hybrid systems and present a generalized method for their optimization. We outline Hybrid Cost Automata (HCA), an extension to Hybrid Automata, where discrete and continuous cost expressions are added. The class of hybrid systems with known spatial paths is defined in the context of HCA. This type of system is common in industry where for example AGVs transpor...

We propose a flipped-Adversarial AutoEncoder (F-AAE) that simultaneously trains a generative model G that maps an arbitrary latent code distribution to a data distribution and an encoder E that embodies an “inverse mapping” that encodes a data sample into a latent code vector. Unlike previous hybrid approaches that leverage adversarial training criterion in constructing autoencoders, F-AAE mini...

In the conventional hybrid active noise control (CHANC) algorithm, filtered-x least mean square (FXLMS) algorithm is used in both feedforward and feedback structures. However, FXLMS does not process reference signal error before updating weight vector, which leads to inadequate robustness of CHANC impulsive environment. To solve this problem, maximum versoria criterion incorporated into HANC (M...