Development Of A Motorcycle Model Detection System Using Machine Learning And Network Algorithms

Abstract

A network framework is used to identify motorcycle models and perform application based categorization. Additionally, the project categorizes the video data according to manufacturer models and utilities. Over 20,000 pictures were captured and uploaded to the system, creating the database. These numerous motorcycle photographs were used to train the network model for effective detection. This work uses a convolutional neural network, which is an efficient approach for creating classifications. To accomplish the
goal, a CNN- ALGORITHMS model called "resnet34" and an activation function called "Relu" are constructed. Layer detection and characterization are done by Resnet34, a 34- layered CNN- ALGORITHMS network. About 20 motorcycle manufacturers and their
models are intended to be processed by the layer. The method is looped repeatedly until it reaches 85% accuracy in the targeted classification.