Are there any plans for a ROI Pooling layer in a tensor stream to detect an object?

Question

I know this question has been asked several times, but I did not find much on Google, except for a few packages written by several authors. In any case, there is a plan to include the roi union layer (officially) in tensorflow, since it is a vital component for detecting an object and other tasks and does not have access to it, it is a pain when using tensor flow.

Any comments or alternative implementation (if verified) are welcome.

Answer 1

I was able to find the answer to my question in the above article. You can use the tf.image.crop_and_resize function to crop any part of the network and resize it. Like ROI combining, you can crop the bounding box (reduce it by the number of downsampling steps, e.g. 32 in VGG16) and resize it to NxN (e.g. 7x7 in VGG16), which can then be fed to the Fully Connected level.