https://github.com/sohhamseal/traffic-sign-classification
Traffic Sign Recognition using CNNs: A comparative approach
https://github.com/sohhamseal/traffic-sign-classification
convolutional-neural-networks machine-learning traffic-sign-classification
Last synced: 7 months ago
JSON representation
Traffic Sign Recognition using CNNs: A comparative approach
- Host: GitHub
- URL: https://github.com/sohhamseal/traffic-sign-classification
- Owner: SohhamSeal
- Created: 2023-04-30T05:14:07.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-04-30T06:50:42.000Z (over 2 years ago)
- Last Synced: 2025-01-13T19:39:52.127Z (9 months ago)
- Topics: convolutional-neural-networks, machine-learning, traffic-sign-classification
- Language: Jupyter Notebook
- Homepage:
- Size: 524 KB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Traffic-Sign-Classification
## Introduction: Traffic Sign Recognition CNN Model
The Traffic Sign Recognition CNN Model is a cutting-edge technology that can help drivers avoid accidents on Indian highways. By recognizing traffic signs with high accuracy, this model ensures that drivers can stay safe even in adverse weather conditions.
## Optimization of the CNN Model
To bring you the best possible results, our team has analyzed various optimizers and compared the architectures of 3, 4, and 6-layered CNN models. We have also tested the impact of optimizers and the number of training epochs to create the most accurate and reliable model for recognizing and classifying traffic signals.
## Powerful GPU and Relevant Keywords
To ensure the model's speed and reliability, we used the NVIDIA Tesla K80 GPU provided by Google Colaboratory's online services. Our model is built using keywords such as Traffic Signs, Convolutional Neural Network, Traffic Sign Recognition, Keras, Google Colab, and Indian highway traffic signs, making it the real deal.
## Importance of Traffic Sign Recognition
Road accidents are a major concern in India, and the driver's failure to read prohibitory traffic signs is a leading cause. Our Driver Support Systems (DDS) are designed to be trustworthy and quick in identifying traffic signs, assisting the driver while traveling, and reducing the risk of accidents.
## Advanced Features of Traffic Sign Recognition Model
Our Traffic Sign Recognition model is designed to take into account the driver's visual system. It is trained to detect and highlight signs that the driver may have missed, while avoiding presenting all of them to reduce confusion. It also keeps track of the driver's field of view, warning them of any signals that are outside of their line of sight or peripheral vision.
## Traffic Accident Statistics
According to the Ministry of Road Transport and Highways' Transport Research Wing study, in 2019, there were more than 4.4 lakh traffic accidents in India, with highways accounting for about 30.5% of those incidents. Even though the nationwide lockdown led to a decrease in traffic accidents to 3.6 lakh in 2020, the percentage of highway accidents maintained at about 31.8% of the total.
## Conclusion: Try Traffic Sign Recognition CNN Model
Our Traffic Sign Recognition CNN Model is a reliable and efficient tool for keeping drivers safe on Indian highways. With its advanced features and optimized architecture, it can help reduce the number of accidents caused by a lack of awareness of traffic signs. This repository proposes four CNN models with autonomous feature generation and output prediction to recognize Indian highway traffic signs. These models showcase the potential of CNNs in recognizing traffic signs, with a classification accuracy of 82.0% on the custom traffic sign dataset. By comparing the effect of different optimizers and epochs, the repository offers insights into improving model performance. However, creating a large-scale training dataset for Indian traffic signs remains a challenge due to uneven distribution. Therefore, more in-depth research and fieldwork are needed to gather a larger dataset that can increase the precision and accuracy of these models, making them more dependable for driver assistance systems.