Faster-rcnn.pytorch: Getting features for each image region
Hello,
is it possible to use this implementation to get features for each image region like they do in the bottom-up attention model of Bottom-Up and Top-Down Attention for Image Captioning and Visual Question Answering?
Thank you!
claudiogreco
All 9 comments
@claudiogreco, Yes, we have that locally. We will add feature extraction part to this repo.
jwyang
on 18 Jan 2018
Thank you! It would be very useful for my work! Do you roughly know when you will commit this feature?
claudiogreco
on 18 Jan 2018
@claudiogreco ， it will be very soon, tomorrow or the day after it.
jwyang
on 18 Jan 2018
Hi @jwyang ,
do you have news about this feature?
Thank you!
claudiogreco
on 22 Jan 2018
Hi @claudiogreco I'll try to upload the feature extraction part in 1 or 2 days.
jiasenlu
on 3 Mar 2018
@jiasenlu - any news on this?
dotannn
on 14 Mar 2018
@jiasenlu any new updates?
ObadaJabassini
on 23 Mar 2018
hey - I've been able to amend the code for extracting the features pretty easily
in the class _fasterRCNN forward method, return also the pooled_feat parameter - it contains feature vector of each region proposal.
return statement of _fasterRCNN.forward() :
return rois, cls_prob, bbox_pred, rpn_loss_cls, rpn_loss_bbox, RCNN_loss_cls, RCNN_loss_bbox,
rois_label, **pooled_feat**
dotannn
on 23 Mar 2018
Closing this issue for now, thanks for @dotannn !
jwyang
on 20 Apr 2018
Related issues
twangnh
·
5Comments
EmmaSRH
·
4Comments
gullalc
·
4Comments
kebijuelun
·
5Comments
gayathrimahalingam
·
3Comments
Most helpful comment
hey - I've been able to amend the code for extracting the features pretty easily
in the class _fasterRCNN forward method, return also the pooled_feat parameter - it contains feature vector of each region proposal.
return statement of _fasterRCNN.forward() :
return rois, cls_prob, bbox_pred, rpn_loss_cls, rpn_loss_bbox, RCNN_loss_cls, RCNN_loss_bbox, rois_label, **pooled_feat**