How to segment images and pass features into neural network using tensorflow?

Question

I am currently working on an image classification project involving Tensorflow and Keras. I am interested in doing image segmentation to remove unnecessary background noise