Vaizdų klasifikavimas pagal suspaudimo algoritmo poveikį jų kokybei
Classification of images according their quality after compression
Author(s): Jevgenij Tichonov, Olga Kurasova, Ernestas FilatovasSubject(s): Visual Arts, ICT Information and Communications Technologies
Published by: Vilniaus Universiteto Leidykla
Keywords: image classification; image quality; image compression; SSIM index method; the classification accuracy of edge detection filters;
Summary/Abstract: In this paper we have investigated an image classification problem where we aim to classify images according their quality after compression. Measures for image quality estimation and methods of image classification have been discussed. The aim of experimental investigation is to determine which features describing the original images have to be used in the classification of the images into two classes with good and bad quality after compression. Quality of images after compression is estimated by a structural similarity (SSIM) index method. We conclude that when using the support vector machine the most accurate classification results have been obtained with preprocessed images by edge detection filters.
Journal: Informacijos mokslai
- Issue Year: 2015
- Issue No: 73
- Page Range: 127-134
- Page Count: 8
- Language: Lithuanian