A problem with fully connected neural networks is their size. When it comes to the practical implementation of neural networks size becomes an important factor. In hardware implementation this becomes a problem of scaling and the number of components required. In the case of computer simulation the problem comes with the large computational cost. The larger and more complex the network the longer it takes to train and once trained it takes longer for the network to perform its recognition task. For this work the concern is with the computational cost.
Feasibility is another important aspect in choosing a neural network thesis subject. Feasibility is a property of a concept that allows you to apply the most appropriate methods for research. This way, you can make sure that all your research results will produce credible results. How can we check if the topic is feasible? A feasible topic has available materials, can be applied with research methods, can produce great results and is timely.
neural network Thesis in Mohali, Chandigarh - Image 1
M. A. Tehrani, R. P. Kleihorst, P. B. L. Meijer and L. Spaanenburg, ``Abnormal Motion Detection in a Real-Time Smart Camera System,'' Third ACM/IEEE International Conference on Distributed Smart Cameras(ICDSC 2009), August 30 - September 2, 2009, Como, Italy. .W.H.A. Schilders, P.B.L. Meijer and E. Ciggaar, ``Behavioural modelling using the MOESP algorithm, dynamic neural networksand the BartelsStewart algorithm,'' Applied Numerical Mathematics, Vol. 58, No. 12, December 2008, pp. 1972-1993. .