Generation of a FIR filter by means of a neural network for improvement of the digital images obtained using the acquisition equipment based on the low quality CCD structure

  • Authors:

  • Jakub Pęksiński;Grzegorz Mikołajczak

  • Affiliations:

  • West Pomeranian University of Technology, Institute of Circuit Theory and Telecommunication Systems, Szczecin, Poland;West Pomeranian University of Technology, Institute of Circuit Theory and Telecommunication Systems, Szczecin, Poland

  • Venue:
  • ACIIDS'10 Proceedings of the Second international conference on Intelligent information and database systems: Part I
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

The paper features the method of application of a neural network for improving the quality of the digital images generated by means of devices for backup and processing of data into a digital form which construction is based on the Charge Coupled Device (CCD) structure. In order to introduce the problem, the digital images were generated by means of two scanners (including a high class and a low class scanner) and afterwards the images were subject to an objective and a subjective evaluation. An objective evaluation was performed using two quality criteria, i.e. MSE (Mean Square Error) and Universal Image Quality Index. A FIR (Finite Impulse Response) filter applied for filtration of a low quality image was obtained as the result of the neural network learning process. The image so generated as the result of filtration features a superior quality in comparison to the original.