• Stars
    star
    131
  • Rank 275,867 (Top 6 %)
  • Language
    Jupyter Notebook
  • Created over 5 years ago
  • Updated over 5 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

An image dataset for training fire and frame detection AI

Fire-Flame-Dataset


An image dataset for training fire and frame detection AI


Fire-Flame-Dataset is a dataset collected in order to train machine learning model to recognize Fire, smoke, and neutral(images without fire or smoke).This a dataset containing about 3000 images and 3 classes which include:

  • Fire
  • Smoke
  • neutral

There are 1000 images in each category and 900 for train and 100 for testing

Download, Training and Prediction


The Fire-Flame-Dataset is provided for download in the release section of this repository. You can download the dataset via this link Fire-Flame-Dataset.

The implementation code in which the model was train with has been provide in this repository. The model was trained with train with resnet50 and a accuracy of 85% on the test data was achieved. The python codebase is contained in fire_flame.ipynb.

Some of the prediction results are shown below:
fire_1

('Image of:', 'Class: Fire', 'Confidence score: 1.0')

fire_2

('Image of:', 'Class: Fire', 'Confidence score: 0.990234375')

neutral_1

('Image of:', 'Class: Neutral', 'Confidence score: 0.99365234375')

neutral_2

('Image of:', 'Class: Neutral', 'Confidence score: 1.0')

smoke_1

('Image of:', 'Class: Smoke', 'Confidence score: 0.4462890625')

smoke_1

('Image of:', 'Class: Smoke', 'Confidence score: 0.9970703125')

Reqirements


  • Python 3
  • Pytorch
  • Numpy
  • Matplotlib
  • TorchFussion

References