search

Dodant / anpr-with-yolo-v4

Automatic License Plate Recognition using Yolo v4 (2020-1 CNU SW Capstone Design Project)
MIT License
39 stars 27 forks source link

Inquiry About the Training Dataset for ALPR with YOLOv4 #3

Open yihong1120 opened 11 months ago

yihong1120 commented 11 months ago

Hello,

I recently came across your project "ALPR with YOLOv4" on GitHub and I'm very impressed with its capabilities and the implementation details provided. The system seems highly efficient for real-time vehicle license plate detection and recognition.

I am particularly interested in understanding more about the training dataset that was used for this project. Specifically, I have the following queries:

  1. Dataset Source: Could you provide more information about the source of the dataset used for training the ALPR system? Is it publicly available or was it a privately curated dataset?

  2. Dataset Details: What are the characteristics of the dataset? For example, the number of images, variety of license plates, image resolutions, diversity in backgrounds, and lighting conditions.

  3. Accessing the Dataset: If the dataset is publicly available, could you kindly share the method to access or download it? In case it's a private dataset, any insights into how it was compiled would be beneficial.

Understanding the dataset will greatly help in comprehending the robustness and the limitations of the ALPR system. Additionally, this information could be beneficial for those looking to conduct further research or development in similar areas.

Thank you for your time and effort in developing this impressive project. Looking forward to your response.

Best regards, yihong1120

Dodant commented 11 months ago

Thank you for visiting my repo.

Dataset was crawled from the web (ex. used car trading sites). and I did labeling datasets on my own. I guess this dataset somehow unavailable. But I can give you some info.

number of images: 2,600 image resolution: over 1080P lighting conditions: mostly daytime, so it's bright variety of LP: old ones (green), new ones (white), taxi (yellow)

It's not a private but I can't find it.. I'm so sorry.