I am doing a deep learning project in which i have to identify different models of cars. i am a bit confused for the following reason:

  • first is what algorithm should I use. I have studied that RCNNs are good algorithms that might work with my project but they are slow; will a faster RCNN reduce accuracy?

  • next is the training data set. I am going to use my own data set which is about 200 images for each model. Is this data sufficient
    for training? Also, what should I know/how should I label and preprocess the data for training?

  • $\begingroup$ I am afraid that the way this question is phrased is too broad. There are a few other deciding factors in this: Why RCNN in the first place? What are the images of? (Front, top, 3/4 view?), What are the inputs and outputs of the system? The "Is the data sufficient for training" depends on the nature of the data itself, so you might have to run some experiments. Have a go at thinking a bit more about the problem and try to edit this question, I think it would help you get more accurate answers. $\endgroup$ – A_A Feb 18 at 8:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.