proposal_layer.py层解读

proposal_layer层是利用训练好的rpn网络来生成region proposal供fast rcnn使用。

proposal_layer整个处理过程：1.生成所有的anchor，对anchor进行4个坐标变换生成新的坐标变成proposals（按照老方法先在最后一层feature map的每个像素点上滑动生成所有的anchor，然后将所有的anchor坐标乘以16，即映射到原图就得到所有的region proposal，接着再用boundingbox regression对每个region proposal进行坐标变换生成更优的region proposal坐标，也是最终的region proposal坐标）　　2.处理掉所有坐标超过了图像边界的proposal　　3.处理掉所有长度宽度小于min_size的proposal　　4.把所有的proposal按score高低进行排序　　5.选择得分前pre_nms_topN的proposal，这是在进行nms前进行一次选择　　6.进行nms处理　　7.选择得分前post_nms_topN的proposal，这是在进行nms后进行的一次选择　　最终就得到了需要传入fast rcnn网络的region proposal。

# --------------------------------------------------------
# Faster R-CNN
# Copyright (c) 2015 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Ross Girshick and Sean Bell
# --------------------------------------------------------

import caffe
import numpy as np
import yaml
from fast_rcnn.config import cfg
from generate_anchors import generate_anchors
from fast_rcnn.bbox_transform import bbox_transform_inv, clip_boxes
from fast_rcnn.nms_wrapper import nms

DEBUG = False

class ProposalLayer(caffe.Layer):
    """
    Outputs object detection proposals by applying estimated bounding-box
    transformations to a set of regular boxes (called "anchors").
    """

    def setup(self, bottom, top):
        # parse the layer parameter string, which must be valid YAML
        layer_params = yaml.load(self.param_str_)

        self._feat_stride = layer_params['feat_stride']
        anchor_scales = layer_params.get('scales', (8, 16, 32))
        self._anchors = generate_anchors(scales=np.array(anchor_scales))
        self._num_anchors = self._anchors.shape[0]

        if DEBUG:
            print 'feat_stride: {}'.format(self._feat_stride)
            print 'anchors:'
            print self._anchors

        # rois blob: holds R regions of interest, each is a 5-tuple
        # (n, x1, y1, x2, y2) specifying an image batch index n and a
        # rectangle (x1, y1, x2, y2)
        top[0].reshape(1, 5)

        # scores blob: holds scores for R regions of interest
        if len(top) > 1:
            top[1].reshape(1, 1, 1, 1)

    def forward(self, bottom, top):
        # Algorithm:
        #
        # for each (H, W) location i
        #   generate A anchor boxes centered on cell i
        #   apply predicted bbox deltas at cell i to each of the A anchors
        # clip predicted boxes to image
        # remove predicted boxes with either height or width < threshold
        # sort all (proposal, score) pairs by score from highest to lowest
        # take top pre_nms_topN proposals before NMS
        # apply NMS with threshold 0.7 to remaining proposals
        # take after_nms_topN proposals after NMS
        # return the top proposals (-> RoIs top, scores top)

        assert bottom[0].data.shape[0] == 1, \
            'Only single item batches are supported'

        cfg_key = str(self.phase) # either 'TRAIN' or 'TEST'
        pre_nms_topN  = cfg[cfg_key].RPN_PRE_NMS_TOP_N　　　　　　　　 #这是在进行nms处理前，从anchor中筛选出前topn个
        post_nms_topN = cfg[cfg_key].RPN_POST_NMS_TOP_N　　　　　　　　#这是经过nms处理后，从anchor中筛选出钱topn个
        nms_thresh    = cfg[cfg_key].RPN_NMS_THRESH
        min_size      = cfg[cfg_key].RPN_MIN_SIZE

        # the first set of _num_anchors channels are bg probs
        # the second set are the fg probs, which we want
        scores = bottom[0].data[:, self._num_anchors:, :, :]
        bbox_deltas = bottom[1].data　　　　　　　　　　　　　　　　　　　　#和anchor_target_layer层一样，获得训练得到4个变化值
        im_info = bottom[2].data[0, :]

        if DEBUG:
            print 'im_size: ({}, {})'.format(im_info[0], im_info[1])
            print 'scale: {}'.format(im_info[2])

        # 1. Generate proposals from bbox deltas and shifted anchors
        height, width = scores.shape[-2:]　　　　　　　　　　　　　　　　　　#这里和anchor_target_layer层一样，都是通过rpn_cls_score得到最后一层特征提取层的长度和宽度if DEBUG:
            print 'score map size: {}'.format(scores.shape)

        # Enumerate all shifts
        shift_x = np.arange(0, width) * self._feat_stride
        shift_y = np.arange(0, height) * self._feat_stride
        shift_x, shift_y = np.meshgrid(shift_x, shift_y)
        shifts = np.vstack((shift_x.ravel(), shift_y.ravel(),
                            shift_x.ravel(), shift_y.ravel())).transpose()

        # Enumerate all shifted anchors:
        #
        # add A anchors (1, A, 4) to
        # cell K shifts (K, 1, 4) to get
        # shift anchors (K, A, 4)
        # reshape to (K*A, 4) shifted anchors
        A = self._num_anchors
        K = shifts.shape[0]
        anchors = self._anchors.reshape((1, A, 4)) + \
                  shifts.reshape((1, K, 4)).transpose((1, 0, 2))
        anchors = anchors.reshape((K * A, 4))　　　　　　　　　　　　　　　　　　　　　　　　#和anchor_target_layer层一样，得到所有的anchor坐标值，并且形状是4列多行　

        # Transpose and reshape predicted bbox transformations to get them
        # into the same order as the anchors:
        #
        # bbox deltas will be (1, 4 * A, H, W) format
        # transpose to (1, H, W, 4 * A)
        # reshape to (1 * H * W * A, 4) where rows are ordered by (h, w, a)
        # in slowest to fastest order
        bbox_deltas = bbox_deltas.transpose((0, 2, 3, 1)).reshape((-1, 4))　　　　　#将bbox_deltas的shape改成和anchors一样，方便下面运算

        # Same story for the scores:
        #
        # scores are (1, A, H, W) format
        # transpose to (1, H, W, A)
        # reshape to (1 * H * W * A, 1) where rows are ordered by (h, w, a)
        scores = scores.transpose((0, 2, 3, 1)).reshape((-1, 1))　　　　　　　　　　　#将scores的shape也变成4列多行

        # Convert anchors into proposals via bbox transformations
        proposals = bbox_transform_inv(anchors, bbox_deltas)　　　　　　　　　　　　　 #通过bbox_deltas将anchors转成proposals，

        # 2. clip predicted boxes to image
        proposals = clip_boxes(proposals, im_info[:2])

        # 3. remove predicted boxes with either height or width < threshold
        # (NOTE: convert min_size to input image scale stored in im_info[2])
        keep = _filter_boxes(proposals, min_size * im_info[2])
        proposals = proposals[keep, :]
        scores = scores[keep]

        # 4. sort all (proposal, score) pairs by score from highest to lowest
        # 5. take top pre_nms_topN (e.g. 6000)
        order = scores.ravel().argsort()[::-1]
        if pre_nms_topN > 0:
            order = order[:pre_nms_topN]
        proposals = proposals[order, :]
        scores = scores[order]

        # 6. apply nms (e.g. threshold = 0.7)
        # 7. take after_nms_topN (e.g. 300)
        # 8. return the top proposals (-> RoIs top)
        keep = nms(np.hstack((proposals, scores)), nms_thresh)
        if post_nms_topN > 0:
            keep = keep[:post_nms_topN]
        proposals = proposals[keep, :]
        scores = scores[keep]

        # Output rois blob
        # Our RPN implementation only supports a single input image, so all
        # batch inds are 0
        batch_inds = np.zeros((proposals.shape[0], 1), dtype=np.float32)
        blob = np.hstack((batch_inds, proposals.astype(np.float32, copy=False)))
        top[0].reshape(*(blob.shape))
        top[0].data[...] = blob

        # [Optional] output scores blob
        if len(top) > 1:
            top[1].reshape(*(scores.shape))
            top[1].data[...] = scores

    def backward(self, top, propagate_down, bottom):
        """This layer does not propagate gradients."""
        pass

    def reshape(self, bottom, top):
        """Reshaping happens during the call to forward."""
        pass

def _filter_boxes(boxes, min_size):
    """Remove all boxes with any side smaller than min_size."""
    ws = boxes[:, 2] - boxes[:, 0] + 1
    hs = boxes[:, 3] - boxes[:, 1] + 1
    keep = np.where((ws >= min_size) & (hs >= min_size))[0]
    return keep

这是这一层的prototxt

layer {
  name: 'proposal'
  type: 'Python'
  bottom: 'rpn_cls_prob_reshape'
  bottom: 'rpn_bbox_pred'
  bottom: 'im_info'
  top: 'rois'
  top: 'scores'
  python_param {
    module: 'rpn.proposal_layer'
    layer: 'ProposalLayer'
    param_str: "'feat_stride': 16"
  }
}

可以看到，bottom[1]就是rpn_bbox_pred

所以上面代码中的bbox_deltas = bottom[1].data就是训练得到的坐标的4个变化值。因为训练rpn网络，本身训练的就是这4个变化值，而不是直接的4个坐标值。

 # bbox deltas will be (1, 4 * A, H, W) format
        # transpose to (1, H, W, 4 * A)
        # reshape to (1 * H * W * A, 4) where rows are ordered by (h, w, a)
        # in slowest to fastest order
        bbox_deltas = bbox_deltas.transpose((0, 2, 3, 1)).reshape((-1, 4))

代码中的这一部分必须理解一下。实际上，bbox deltas，也就是要学习的那4个变换值。首先必须知道的是，这4个变换值是训练学习来的，是由卷积训练来的，来自于rpn_bbox_pred这一层，他是一个feature map， shape是（4×anchor个数，h，w）。如何将这个feature map和生成的anchor进行变换，首先必须shape一样才能加或者其他运算。所以，这里所做的就是将bbox deltas的shape变成了和anchors一样的4列多行，4列就代表着x,y,w,h。

注意：无论是anchors还是bbox deltas，还是scores，他们的shape都是多行4列，排列的顺序都是（h，w，a），即第一行是h，w，a，第二行是h+1，w，a，当h排完了，再排w的变换，最后才是a