45

Has anyone implement the FRCNN for TensorFlow version? I found some related repos as following:

  1. Implement roi pool layer
  2. Implement fast RCNN based on py-faster-rcnn repo

but for 1: assume the roi pooling layer works (I haven't tried), and there are something need to be implemented as following:

  • ROI data layer e.g. roidb.
  • Linear Regression e.g. SmoothL1Loss
  • ROI pool layer post-processing for end-to-end training which should convert the ROI pooling layer's results to feed into CNN for classifier.

For 2: em...., it seems based on py-faster-rcnn which based on Caffe to prepared pre-processing (e.g. roidb) and feed data into Tensorflow to train the model, it seems weird, so I may not tried it.

So what I want to know is that, will Tensorflow support Faster RCNN in the future?. If not, do I have any mis-understand which mentioned above? or has any repo or someone support that?

RyanLiu
  • 1,557
  • 2
  • 18
  • 27
  • SmoothL1Loss should be relatively easy to implement using the actual tf for ROI pooling no idea... – jeandut Jul 13 '16 at 11:56
  • I am working on the similar target of your question. I found that it hard to represent dynamic bboxes in tensor. That maybe the reason why the method 2 you mentioned use caffe to pre-process data. I am trying to figure out whether there is some other way to achieve that in TensorFlow. – Da Tong Sep 23 '16 at 06:01
  • 4
    How about [this implementation](https://github.com/smallcorgi/Faster-RCNN_TF)? – Shai Jan 09 '17 at 06:05

1 Answers1

16

Tensorflow has just released an official Object Detection API here, that can be used for instance with their various slim models.

This API contains implementation of various Pipelines for Object Detection, including popular Faster RCNN, with their pre-trained models as well.

Wahaj
  • 95
  • 7
gdelab
  • 6,124
  • 2
  • 26
  • 59