Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Using Neural Networks in Forensic Writer Identification

Keywords:

D. A. Nikonets


The paper deals with the application of the neural networks for the solution of the problems of the forensic handwriting expertise. The primary goals of the forensic handwriting expertise and possible ways of their solving are examined. Existing databases of the hand-written samples used as training and control sets for training of neural networks at the solving of problems of the off-line recognition of the hand-written text and identification of the executor are described. Comparison of various standards of the description of characteristics of handwriting objects for identification of the executor, taken from the image of the hand-written text is presented. Modern software for forensic identification of the executor of the hand-written documents is examined. Accuracy of identification of the executor, received at use of these programs, and, also, filters of preliminary processing of the image of documents applied in them, neural networks and characteristics of the hand-written text are briefly described. Short data on some neural network programs which can be used for the solving of forensic problems by means of neural networks are listed. Authors in detail describe formation of a reference database of handwriting indications for capital and lower case letters of the Russian language. The elaboration of the program for forensic handwriting expert «FHWE v.1.0» and application of the recognition subsystem «RSFHWE v.1.0» for identification of the executor of the hand-written documents are described
References:

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio