Open AccessLow-loss image compression techniques for cutting tool images: a comparative study of compression quality measures
Author(s) -
Fábio Henrique Pereira,
Elesandro Antônio Baptista,
Nivaldo Lemos Coppini,
Rafael Do Espírito-Santo,
Ademir João de Oliveira
Publication year - 2010
Publication title -
exacta
Language(s) - English
Resource type - Journals
eISSN - 1983-9308
pISSN - 1678-5428
DOI - 10.5585/exacta.v8i2.2000
Subject(s) - image compression , compression (physics) , computer science , data compression , data compression ratio , artificial intelligence , peak signal to noise ratio , mean squared error , matlab , principal component analysis , image quality , compression ratio , computer vision , pattern recognition (psychology) , image (mathematics) , mathematics , image processing , statistics , engineering , materials science , internal combustion engine , automotive engineering , composite material , operating system
This work accomplishes a comparative study between two distinct image compression techniques, namely the Lifting technique and the Principal Components Analysis (PCA), in order to determine what of these two approaches is more appropriate for cutting tool wear images analysis. Lifting and Principal Components Analysis were applied in original images of a cutting tool for producing a low resolution version, while keeping the more important details of the image. The low-loss image compression quality provided by these techniques was expressed in terms of the compression factor (ρ), the Mean Square Error (MSE) and the Peak Signal-to-Noise Rate (PSNR) provided by the image compression process. The tests were accomplished using the high-performance language for technical computing MATLAB®, and the results shown that the PCA technique presented the best values of PSNR with low compression rates. However, with high values of compression rates the lifting technique gave the highest PSNR.