Revision history [back]

So what I suggest is an approach on the Darknet19 CNN classifier.

• Use transfer learning to train a new model using the original architecture
• Use OpenCV DNN interface to read in the model and process your images

I am pretty sure collecting 50 sample patches will already get you pretty far!

So what I suggest is an approach on the Darknet19 CNN classifier.CNN classifier. If time is not an issue, than go for the DenseNet201 architecture, because it is simply so powerful!

• Use transfer learning to train a new model using the original architecture
• Use OpenCV DNN interface to read in the model and process your images

I am pretty sure collecting 50 sample patches will already get you pretty far!