I wish to implement Panoptic FPN (which adds an extra branch to Mask R-CNN for semantic segmentation) using TensorFlow, and use pre-trained models for its ResNet-FPN backbone. However, I am add a crossroads: do I use TensorFlow's Keras API, or the TensorFlow Object Detection API? Is the latter even capable of semantic segmentation? What do you recommend?

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.