Skip to content

Simple Improved LeNet-5 on GTSRB (Traffic Sign Classification)

Notifications You must be signed in to change notification settings

ElementQi/Improved-LeNet-5-on-GTSRB

Repository files navigation

Simple Improved LeNet-5 on GTSRB (Traffic Sign Classification)

WARNING

There are DNS spoofing in some regions in China or other places, which leads to the images cannot be shown.

Declaration

This is the project for class Data Science Programming II.

There are report file, presentation ppt and draft, jupyter notebook file(s), log files, model(HDF5) files, and dataset examples in the repository.

Reference

Mainly based on the paper.

Zaibi, Ameur & Anis, Ladgham & Sakly, Anis. (2021). A Lightweight Model for Traffic Sign Classification Based on Enhanced LeNet-5 Network. Journal of Sensors. 2021. 10.1155/2021/8870529.

The model conception map:

DNS spoofing?

Dataset

43 classes for Traffic Sign Classification

DNS spoofing?

I download those:

  • GTSRB_Final_Test_GT.zip (csv) (I unzip this file into the test images folder)
  • GTSRB_Final_Test_Images.zip (ppm+csv)
  • GTSRB_Final_Training_Images.zip (ppm+csv)

There is also a file called signnames.csv, which maps numbers into class names. I found it on kaggle.

Project in One Figure

DNS spoofing?

About

Simple Improved LeNet-5 on GTSRB (Traffic Sign Classification)

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published