Viet-Anh on Software Logo

What is: Precise RoI Pooling?

SourceAcquisition of Localization Confidence for Accurate Object Detection
Year2000
Data SourceCC BY-SA - https://paperswithcode.com

Precise RoI Pooling, or PrRoI Pooling, is a region of interest feature extractor that avoids any quantization of coordinates and has a continuous gradient on bounding box coordinates. Given the feature map F\mathcal{F} before RoI/PrRoI Pooling (eg from Conv4 in ResNet-50), let wi,jw_{i,j} be the feature at one discrete location (i,j)(i,j) on the feature map. Using bilinear interpolation, the discrete feature map can be considered continuous at any continuous coordinates (x,y)(x,y):

f(x,y)=i,jIC(x,y,i,j)×wi,j,f(x,y) = \sum_{i,j}IC(x,y,i,j) \times w_{i,j},

where IC(x,y,i,j)=max(0,1xi)×max(0,1yj)IC(x,y,i,j) = max(0,1-|x-i|)\times max(0,1-|y-j|) is the interpolation coefficient. Then denote a bin of a RoI as bin={(x1,y1),(x2,y2)}bin=\{(x_1,y_1),(x_2,y_2)\}, where (x1,y1)(x_1,y_1) and (x2,y2)(x_2,y_2) are the continuous coordinates of the top-left and bottom-right points, respectively. We perform pooling (e.g. average pooling) given binbin and feature map F\mathcal{F} by computing a two-order integral: