jsk_recognition_utils.chainermodels.roi_pooling_2d Namespace Reference

Classes
class	ROIPooling2D

Functions
def	_roi_pooling_slice (size, stride, max_size, roi_offset)
def	roi_pooling_2d (x, rois, outh, outw, spatial_scale)

Function Documentation

def jsk_recognition_utils.chainermodels.roi_pooling_2d._roi_pooling_slice (   size,   stride,   max_size,   roi_offset  )

private

Definition at line 39 of file roi_pooling_2d.py.

def jsk_recognition_utils.chainermodels.roi_pooling_2d.roi_pooling_2d	(		x,
			rois,
			outh,
			outw,
			spatial_scale
	)

Spatial Region of Interest (ROI) pooling function.

This function acts similarly to :class:`~functions.MaxPooling2D`, but
it computes the maximum of input spatial patch for each channel
with the region of interest.

Args:
    x (~chainer.Variable): Input variable.
    rois (~chainer.Variable): Input roi variable.
    outh (int): Height of output image after pooled.
    outw (int): Width of output image after pooled.
    spatial_scale (float): Scale of the roi is resized.

Returns:
    ~chainer.Variable: Ouptut variable.

See the original paper proposing ROIPooling:
`Fast R-CNN <http://arxiv.org/abs/1504.08083>`_.

Definition at line 313 of file roi_pooling_2d.py.