gaiasd/DFireDataset

Pre-trained smoke detection models?

Closed this issue · 6 comments

Intersting repo. Your sample image shows smoke + fire detection, but there is no pre-trained models available. Can you please share the that model produced that? I'm specifically interested in smoke part.

gaiasd commented

We are glad you are interested! In this repository you will find links to models, but unfortunately we cannot publicly distribute trained weights because of copyrights.

Thanks but that is only fire and not smoke.

@arianaa30 Thank you for your interest in our work. In fact, these models detect fire and smoke. The D-Fire dataset includes both classes and the models were trained on it.

Here is an example of smoke detection from one of our models:

CADFire005-output

You can see more results of these models in our following papers:

Good. I'll try them to get smoke! I throught they are trained only on the fire images of the Dfire dataset

@arianaa30 Excellent! I will close this issue. Any other problem, feel free to open a new issue. If the models and the dataset are useful to you and you want to help us spread the word about our work, please star our repositories. Thank you!

Hi,
I was trying the models from the other repo you mentioned. Does it belong to you?
I have a few questions. Would you be able to answer?

I am running the model with python3 baseline.py --video ex1.mp4 --model mobilenet, but it only shows some bash outputs and no output videos. Does it save the output file somewhere?

1/1 [==============================] - 0s 25ms/step
Time taken =  0.04462027549743652
Prediction: non_fire
1/1 [==============================] - 0s 25ms/step
Time taken =  0.043842315673828125
Prediction: non_fire
1/1 [==============================] - 0s 26ms/step
Time taken =  0.04700660705566406
Prediction: non_fire
1/1 [==============================] - 0s 27ms/step
Time taken =  0.050565481185913086
Prediction: non_fire
Results
     video    network  detected  first_frame  time_avg
0  ex1.mp4  mobilenet      True           10  0.049584