Title :
A new dynamic strategy of Recurrent Neural Network
Author_Institution :
Sch. of Comput. Sci. & Eng., Univ. of New South Wales, Sydney, NSW, Australia
Abstract :
Recurrent neural networks are widely used in many applications that need to store and update context information. The number of recurrent cycle is very important to many tasks in terms of accuracy and computation time. In this paper, a dynamic strategy is extended to recurrent internal symmetry neural network, and back propagation is trained for image segmentation tasks.
Keywords :
backpropagation; image segmentation; recurrent neural nets; back propagation; context information; image segmentation; recurrent cycle; recurrent internal symmetry neural network; Application software; Australia; Computer networks; Computer science; Convergence; Equations; Image segmentation; Neural networks; Recurrent neural networks; Testing; back propagation; dynamic Cycle; internal symmetry; recurrent neural networks;
Conference_Titel :
Cognitive Informatics, 2009. ICCI '09. 8th IEEE International Conference on
Conference_Location :
Kowloon, Hong Kong
Print_ISBN :
978-1-4244-4642-1
DOI :
10.1109/COGINF.2009.5250690
