Vehicle Feature Detection And Rating System

Abstract

The aim of our project was to build an innovative system that intends to provide a package of features which can extract a car’s details using its provided images. For the purpose of this project, our targeted domain is the booming online car resale platform. We wanted our project to be able to provide services that will streamline the consumer’s experience while reducing the scope of human error. Our feature will take images as inputs from the user and then using these images will extract the following details: Make, Model, Color and Registration Number of the car. These obtained details will then be displayed to the user who can verify whether they want to continue with these details or would like to modify them. Later on, the user also has the ability to opt for a price estimation using our price approximator feature. This handy feature is formed on the basis of currrent market trends and will be useful for users to acquire a basic approximation of the price that their car might be able to fetch on the market. However, the above feature does not currently take into account the car’s condition and hence can only provide a basic reckoning that the car’s price might be able to fetch. The user can then employ these details in their advertisements, making the process less complex and repetitive and simultaneously enhancing the user experience.