Skip to content

MohithReddy1/Traffic-Sign-Classification

Repository files navigation

Traffic-Sign-Classification

Abstract

Traffic symbols are common in our daily life. It includes information that ensures safety and endless number of people that surround us.

Having the ability to recognize traffic symbols automatically qualifies us to create "more brilliant automobiles."To properly parse and interpret the roadway, self-driving vehicles require traffic sign recognition. To support and assure drivers, "driver alert" frameworks within cars must understand the street around them.

PC vision and profound learning may comprehend a variety of challenges, including traffic sign recognition. The traffic signs are placed as a last-minute addition or at the street's beginning. They provide direction on how to go when out and about, allowing traffic to flow safely and easily. Everyone should be aware of the traffic signs!

Dataset Click here.

Project Objectives

  • To identify the traffic signs and classify them.
  • To classify the detected traffic signs to their specific sub-classes.
  • Save the model with best accuracy
  • Deploy the model to upload and classify the image and show output.

Experimental Results

The CNN model has been trained on a GTSRB dataset that contains train images of 43 classes and test images. The model has been developed for the classification of traffic signs of 43 classes and there by producing an accuracy of 95.27 % Accuracy and 99.044%.

Deployment

We have saved the CNN model. By using tkinter package for creating app like interface using python code with that model that have saved before. In that tkinter app first we need to upload any image of different traffic signs and click on classify image button and it display the result what is that traffic sign of 43 classes.

Output Image1 Output Image2
output1 output2

About

No description or website provided.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published