Acceleration of knee MRI cancellous bone classification on google colaboratory using convolutional neural network

Abstract

The field of Biomechanical engineering and Orth ology is related to knee bone tissue in which cancellous bone lies. The cancellous bone also called spongy bones has greater surface area for this it causes Osteoporosis (OP), Osteoarthritis (OA), and knee cartilage and knee replacement. The knee bone images are measured mostly by Magnetic Resonance Imaging (MRI).In this paper we presented deep learning model on cancellous bones (tiff type) MRI through Convolutional Neural Network (CNN) to predict the image classification which achieved 99.39 % accuracy. The sample size of images are 185 cancellous MRI and 185 tiff images. Further we trained our model on cloud service that is Google Colabaratory (Colab) which is Graphical Processing Unit (GPU). The accuracy of this model is same but the execution time per min decreases on GPU environment. We increased the no of epochs 20 then 50 its execution time is 10 times less than CPU. The execution time on GPU google Colab is 2.23 (mins) and on CPU its 24.23(mins).