Božica Bojović, Image size and sample areas interaction effects at contact lens surface comparison… Contemporary Materials, IV1 (2013) Page 69 of 75 Original scientific papers UDK 681.7+66.017/.018 doi: 10.7251/COMEN1301065B EFFECTS OF IMAGE SIZE AND SAMPLE AREAS ON COMPARISON OF CONTACT LENS SURFACES BASED ON FRACTAL DIMENSION Božica Bojović Faculty of Mechanical Engineering, University of Belgrade, Kraljice Marije 16, 11120 Beograd, Serbia Abstract: Methods used for calculation of fractal dimension demand large image resolution and adequate sample size, in terms of roughness threshold that defines spatial scope for rough surface fractal properties. Imaging device operators, on the one hand, recommend the image size and sample area based on experience and expertise, in order to minimize the imaging time. On the other hand, engineers make decisions based on their own requirements. To overcome these problems, in this paper we used ANOVA statistical approach (one-way and multi comparison) so as to establish significant image size and sample area. The conclusion made in this paper will enable decision guidelines on selection of parameters for new nanophotonic lenses imaging by scanning microscopes in near futu- re. Key words: fractal dimension, imaging, surface roughness, ANOVA. 1. INTRODUCTION Topography of engineering surface created by one or more manufacturing processes can be charac- terized by widely used standard roughness parame- ters, as well as by a single and alternative fractal dimension. Gas-permeable (RGP) micro-machined contact lenses belong to a group of biomedical surfa- ces with finer grade of quality, called optical quality. ISO 1302:2002 Standard defines roughness parame- ters values for surface quality labeled with grade number N1 to N12, with R a values of 0.025μm to 50μm, respectively. Although ISO standard has a certain recommendation for standard roughness parameter R a values from 0.025μm to 0.008μm, there are not the values for sampling length. The fractal dimension is the intrinsic surface property as oppo- sed to any standard roughness parameter that conta- ins only partial information. Fractal dimension is a scale invariant parameter and, therefore is more sui- table for texture characterization, especially in case of biomedical surfaces recorded by scanning microscopy. Methods used for fractal dimension cal- culation demand large image resolution, for example 512x512 pixels. Furthermore, the size of sample previously prepared from machined surface is very important in terms of roughness threshold defining the spatial scope for fractal dimension [1]. There are two practical problems faced by engineers during surface imaging in order to charac- terize machined surface. The first problem is to determine the values of the surface parameters that will characterize the desired intrinsic property and the second is to minimize time for imaging, both in terms of the image size and the sample area that have the main influence on it. The decision made by engineers is based not only on their own requirements but also on imaging device operators’ experience and expertise. To overcome these pro- blems, in this paper we used the well-known ANO- VA test as a statistical approach to establishing sig- nificant image size and sample area. Relationships between various scans taken from different samples made of gas-permeable contact lens basic material are investigated using Matlab. Conclusion will be used for ongoing investigation that considers the new “nano-photonic” material for contact lens that is composed of basic material, doped with three types of nanomaterials. 2. MATERIALS AND METHODS 2.1. Sample Preparation Samples are taken from contact lenses that are manufactured from fluorosilicone acrylate doped * Correspondig author: bbojovic@mas.bg.ac.rs