dear researchers at present we are working in fire detection projcts ,i have a plan to implement the CNN with SVM for classification method. if you hav any best or recent techniques please let me know.
Have a look at::: Article Image fire detection algorithms based on convolutional neural networks
http://www.jcreview.com/fulltext/197-1587743254.pdf
Article An Efficient Deep Learning Algorithm for Fire and Smoke Dete...
As a first step, I suggest using the ResNet models. You can load the weights with the weights trained on ImageNet. If you are using Pytorch with torchvision, this is just a one line code.
To use SVM as a classifier, you can remove the last layer of the ResNet which is a fully-connected layer and use SVM instead.
Here are some links that can help you:
https://pytorch.org/docs/stable/torchvision/models.html
https://www.researchgate.net/post/In_CNN_can_we_replace_fully_connected_layer_with_SVM_as_classifier
Article AUTOMATED SYSTEM-COGNITIVE ANALYSIS OF SPECIFIC SPECTRAL AND...
I recommend you random forest and Backpropagation neural network. These two machine learning method are a good and easy to build especially in matlab but before that you should also choose the appropriate technique for features selection
- Badges
- Science topic
- Similar topics
- Data Mining
- Classification