Dynamic Neural Network Based Human Recognition through Mouse Movement

K. Nagi Reddy; Dr. C. D. Naidu; Dr. K. Suvarchala

In this paper we present a system to identify a person through mouse movement using Hopfield based Dynamic Neural Network. The DNN has a composite structure wherein each node of the network is a Hopfield network by itself. The DNN with reuse is more suitable to accomplish the present task as it is having the properties like associative memory with 100% recall, large storage capacity, avoiding spurious states and converging only to user specified states. Since the DNN with reuse operates with binary data while the information in Mouse Movement Database is in decimal form, the decimal digits are mapped a binary string called descriptor. The network is trained by descriptors (binary strings), which are obtained through the steps indexing through thresholding. These descriptors are used as exemplar patterns to store them in the DNN with reuse by training. The associative memory property of the network makes it to recall a closest pattern of the given query pattern among the memorized patterns. The experimental results are reported to corroborate that high level of precision can be obtained for efficient recall of inexact queries using the DNN with reuse.

Select Volume / Issues:

• Volume 6,Issue 6, Nov. 2015

Year:

2015

Type of Publication:

Article

Keywords:

Hopfield Network; Dynamic Neural Network; Reuse; Associative Memory; Thresholding; Exemplar Patterns

Journal:

IJECCE

Volume:

6

Number:

6

Pages:

752-755

Month:

November