search

Wentong-DST / im2p

Tensorflow implementation of paper: A Hierarchical Approach for Generating Descriptive Image Paragraphs
MIT License
15 stars 5 forks source link

I have a question about the regions detector #3

Closed stoensin closed 6 years ago

stoensin commented 6 years ago

I read the paper and your code. I didn't find the following part of the code ：

4.1. Region Detector

“ These regions are projected onto the convolutional feature map, and the corresponding region of the feature map is reshaped to a fixed size using bilinear interpolation and processed by two fully-connected layers to give a vector of dimension D for each region.”

Please note the bold text section

So I came to ask for help about this qustion. How do you implement it? ：）

stoensin commented 6 years ago

I got it，it's densecap's Recognition Network,