Fakulta elektrotechniky a informatiky / Faculty of Electrical Engineering and Informatics
Stálý URI pro tuto komunituhttps://hdl.handle.net/10195/3847
Práce obhájené před rokem 2008 jsou uloženy pouze v kolekci Vysokoškolské kvalifikační práce
Procházet
5 výsledky
Search Results
Konferenční objektpeer-reviewedpostprint Omezený přístup Neural Network for Smart Adjustment of Industrial Camera - Study of Deployed Application(Springer, 2018) Doležel, Petr; Honc, DanielSince machine vision is gaining more and more interest lately, it is necessary to deal with correct approaches to visual data acquisition in industry. As a particular part of this complex problematics, a technique for the industrial camera exposure time and image sensor gain tuning is presented in this contribution. In comparison to other approaches, a human expert photographer is used instead of explicitly defined cost function. His knowledge is transformed into an artificial expert system represented by a feedforward neural network. The expert system then provides the suitable exposure time and image sensor gain to gather sharp and balanced images.Článekpeer-reviewedpublished Omezený přístup COMPUTATIONALLY SIMPLE NEURAL NETWORK APPROACH TO DETERMINE PIECEWISE-LINEAR DYNAMICAL MODEL(České vysoké učení technické v Praze, 2017) Doležel, Petr; Heckenbergerová, JanaThe article introduces a new technique for nonlinear system modeling. This approach, in comparison to its alternatives, is straight and computationally undemanding. The article employs the fact that once a nonlinear problem is modeled by a piecewise-linear model, it can be solved by many efficient techniques. Thus, the result of introduced technique provides a set of linear equations. Each of the equations is valid in some region of state space and together, they approximate the whole nonlinear problem. The technique is comprehensively described and its advantages are demonstrated on an example.Článekpeer-reviewedpostprint Otevřený přístup Weight Initialization Possibilities for Feedforward Neural Network with Linear Saturated Activation Functions(Elsevier Science BV, 2016) Doležel, Petr; Škrabánek, Pavel; Gago, LumírInitial weight choice is an important aspect of the training mechanism for feedforward neural networks. This paper deals with a particular topology of a feedforward neural network, where symmetric linear saturated activation functions are used in a hidden layer. Training of such a topology is a tricky procedure, since the activation functions are not fully differentiable. Thus, a proper initialization method for that case is even more important, than dealing with neural networks with sigmoid activation functions. Therefore, several initialization possibilities are examined and tested here. As a result, particular initialization methods are recommended for application, according to the class of the task to be solved.Článekpeer-reviewedpublished version Omezený přístup Simulation of predictive control algorithm for rotary furnace producing magnesite sinter(Trans Tech Publications, 2015) Mariška Martin; Taufer Ivan; Koštial Imrich; Doležel Petr; Palička PavolThe paper describes the design of predictive control algorithm for rotary furnace control. At the beginning of the paper, there are defined the main aims of the predictive control. Then, the functional diagram using artificial neural networks as a reference model is proposed and, eventually, the functionality of described approach is demonstrated on a simulated example.Konferenční objektpeer-reviewedpostprint Omezený přístup European Starling Detection in Agricultural Areas Using Multilayer Artificial Neural Network(IEEE (Institute of Electrical and Electronics Engineers), 2015) Doležel Petr; Rozsíval Pavel; Mariška MartinThe use of neural network clasifier to recognize a pest bird in agricultural areas is presented in this contribution. Firstly, the idea of comprehensive system of protection against pest birds is outlined. Then, the method of detection is described, the process of neural network design is illustrated and, in the end, the neural network is validated using data gathered in fields.