Evaluation of Surface Roughness in Additive Manufactured customised implant using Artificial Neural Network based on 2D Fourier transform –A Machine Vision approach

The purpose of this work is to evaluate the surface roughness (Ra) of the customised bone implant fabricated by Selective Laser Sintering (SLS). Computer Tomography (CT) scan data of the femur bone was taken for the study. The scan data was converted to .stl file format. Taguchi’s design of experiments was conducted to fabricate the customised implants using SLS. The quantitative measures of surface roughness are extracted in the spatial frequency domain using a two-dimensional Fourier Transform (FT). Artificial Neural Network (ANN) was trained using feed forward back-propagation algorithm. The surface roughness values obtained by trained ANN using image processing technique and the traditional stylus probe methods are then compared. The comparison results show that the proposed method gives better results on par with the traditional method. For the validation the customised implants with high and low roughness surface were then seeded with 3T3 fibroblast cells and its cell viability was assessed by MTT assay.