Go to Polygence Scholars page
Aadya Vemulapalli's cover illustration
Polygence Scholar2022
Aadya Vemulapalli's profile

Aadya Vemulapalli

Stoller Middle SchoolClass of 2026

About

Projects

  • "How effective are different CNN architectures for creating a bounding box around a road sign and classify it as a traffic light, stop sign, speed limit or crosswalk sign?" with mentor Aditya (Working project)

Project Portfolio

How effective are different CNN architectures for creating a bounding box around a road sign and classify it as a traffic light, stop sign, speed limit or crosswalk sign?

Started June 6, 2022

Abstract or project description

An average of 2.5 million car crashes occur annually from which 1.3 million people die. Drunk drivers are the cause of around 10 thousand deaths each year which is in the United States alone. With the implementation of autonomous cars, roads can be revolutionized, and lives can be spared. Machine learning concepts have started to become widely popular throughout many different uses of technology and are now used in an array of various projects. These concepts can be used in order to create different necessary aspects of autonomous cars ranging from detecting nearby cars to detecting empty parking spaces. Convolutional neural network (CNN) is a machine learning algorithm and can be used to detect road signs such as stop signs, traffic lights, speed limits, etc. This will allow autonomous cars to follow driving rules. The goal of this project is to use CNN to identify road signs using python. The program will take a given image of a road sign and will use CNNs to describe the shown sign. The accuracy rate of this program will be tested in the end. Discussion on the advantages and disadvantages of using CNN to approach this project will also be included.