The Application of the Neural Network for Snakes and Generation of Learning Data

• Iida kentarou

  Minolta corp.

• Imai Masakazu

  Graduate School of Information Science, Nara Institute of Science and Technology

• Uno Takesi

  Graduate School of Information Science, Nara Institute of Science and Technology

Bibliographic Information

Other Title

• スネークスへのニューラルネットワークの適用と学習用データ作成の手法

Abstract

We have to set optimum coefficients of energy tenms in Snakes to obtain desired contours in conver-gence of Snnakes. There is a method to apply neural network that had learned the coefficients for the decision where the nodes of Active Tubes move to in Greedy Algorithm. Active Tubes is accumulation of Snakes along with time axis. As, the learning data were not enough, the convergence results of Snakes by the method is not satisfactory. In this paper, we will propose an extended method to make learning data, for improvement in convergence performance of the Snakes that were applied neural networks. This method will give us more effective learning data from the given teacher contours.

Journal

• IPSJ SIG Notes

  IPSJ SIG Notes 96 (31), 29-36, 1996-03-21

  Information Processing Society of Japan (IPSJ)

Details

• CRID: 1571417127175862400
• NII Article ID: 110002930265
• NII Book ID: AN10100676
• ISSN: 09196072
• Text Lang: ja
• Data Source:
  • CiNii Articles