I am trying to build a system to segment vehicles using a deep convolutional neural network. I am familiar with predicting a set amount of points (i.e. ending a neural architecture with a Dense layer with 4 neurons to predict 2 points(x,y) coords for both). However, vehicles come in many different shapes and sizes and one vehicle may require more segmentation points than another. How can I create a neural network that can have different amounts of output values? I imagine I could use a RNN of some sort but would like a little guidance. Thank you

For example, in the following image the two vehicles have a different number of labeled keypoints.

enter image description here


Landmarks are nice when you have a fixed amount for every image but I don't think it is the right approach for your problem. Instead I think you should look into models that segment images by applying a mask. A place to start could be to look into Mask R-CNN.

Here is the paper: Mask R-CNN

Here is how to train a ready made Mask R-CNN implementation with your own dataset using Keras: Tutorial

  • 1
    $\begingroup$ Didn't think about the problem at a pixel-level. Thanks for the resource! $\endgroup$ – user3647894 Apr 29 at 17:51

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.