Discriminative ROI Pooling

[Kevin43614]

Hello I want to ask one question about your paper.
You say you use a pooling size of 7 × 7 (where k = 7) for classification, so "light-weight offset prediction only requires a k/2 ×k/2
sized RoIAlign" which means pass 3.5*3.5's feature map through fully connected layers ?

[JialeCao001]

@Kevin43614 Thanks for interest. Yeah. I remeber that we use 3x3 for offset prediciton.

[Kevin43614]

@JialeCao001 If the input size used to offset prediction is 3x3 , and through fully connected layers , how to do RoIAlign and generate a 2k2k(1414) size feature map ?

[JialeCao001]

@Kevin43614 After fc layers, we reshape the vector to feature map and upsample the feature map.

[z0978916348]

@JialeCao001
I am also confused about this part.
Can you provide more details about operations from three fc layers to generate (2k x 2k) resolution feature map?

[JialeCao001]

@z0978916348 Please refer the code.

D2Det/mmdet/ops/dcn/deform_pool.py

Line 199 in a76781a

offset = data.new_empty(0)