3D MODEL CLASSIFICATION USING CONVOLUTIONAL NEURAL NETWORK

Abstract

3D modelling is a process ofdeveloping a mathematical representation of any surface of an object. Its capabilities allow you to create 3D images that are as realistic as the actual objects. The objective of this project is to apply the concept of a Deep Learning algorithm namely, Convolutional Neural Networks (CNN) in 3D model classification. This project will focus on applying Neural Network (NN) machine learning methods for 3D model classification. It gave a detailed analysis ofthe process ofCNN algorithm. Then it applied the convolutional neural network to implement the 3D model classification by python. The algorithm is tested on various standard datasets of ShapeNet. The goal is to classify 3D models directly using convolutional neural network. Most of existing approaches rely on a set of human-engineered features. It uses 3D convolutional neural network to let the network learn the features over 3D space to minimize classification error. It trained and tested over ShapeNet dataset with data augmentation by applying random transformations. It made various visual analysis to find out what the network has learned. The performance of the algorithm is evaluated based on the quality metric and classification accuracy. The graphical representation of the experimental results is given based on the number oftraining epochs. The experimental result analysis based on the quality metrics and the graphical representation proves that the algorithm (CNN) gives good classification accuracy for all the tested datasets