2

I'm making a Car Damage Detection model which would have 2 classes to detect upon. My dataset has a total of 300 images (out of which I'd be using some for testing), which are totally insufficient to train the model from scratch.

Can I use a pre-trained model to train my dataset and detect images based on the 2 classes, and if yes, then which one should be the best on my problem set?

P.S. - I would prefer to use Faster-RCNN on my dataset

Preetkaran Singh
  • 123
  • 4

2 Answers2

0

Any pre-trained image recognition model would probably work fine for this purpose. Once they learn main features, they fine tune really well. Take a look at Keras Applications or TensorFlow Hub.

serali
  • 1,212
  • 1
  • 7
  • 19
0

For this task you should use any model pre-trained on the COCO's dataset. We managed to get quite good results with only 80 images with a pre-trained MaskRCNN.

Sebastian Savidan
  • 1