Faster Rcnn 源码解析（三）—— bbox_transform.py

简介：

这个代码里面主要是一些在anchor_targte_layer.py和proposals_layers.py中使用到的一些函数，比较简单，主要是帮助以上两个代码理解。

源码：

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

import numpy as np
#计算与anchor有最大IOU的GT的偏移量
#ex_rois：表示anchor；gt_rois：表示GT
def bbox_transform(ex_rois, gt_rois):
    #得到anchor的（x,y,w,h）
    ex_widths = ex_rois[:, 2] - ex_rois[:, 0] + 1.0
    ex_heights = ex_rois[:, 3] - ex_rois[:, 1] + 1.0
    ex_ctr_x = ex_rois[:, 0] + 0.5 * ex_widths
    ex_ctr_y = ex_rois[:, 1] + 0.5 * ex_heights
    # 得到GT的（x,y,w,h）
    gt_widths = gt_rois[:, 2] - gt_rois[:, 0] + 1.0
    gt_heights = gt_rois[:, 3] - gt_rois[:, 1] + 1.0
    gt_ctr_x = gt_rois[:, 0] + 0.5 * gt_widths
    gt_ctr_y = gt_rois[:, 1] + 0.5 * gt_heights
    #按照损失函数中的计算公式，计算，得到对应的偏移量
    targets_dx = (gt_ctr_x - ex_ctr_x) / ex_widths
    targets_dy = (gt_ctr_y - ex_ctr_y) / ex_heights
    targets_dw = np.log(gt_widths / ex_widths)
    targets_dh = np.log(gt_heights / ex_heights)

    targets = np.vstack(
        (targets_dx, targets_dy, targets_dw, targets_dh)).transpose()
    return targets
#根据anchor和偏移量计算proposals
def bbox_transform_inv(boxes, deltas):
    if boxes.shape[0] == 0:
        return np.zeros((0, deltas.shape[1]), dtype=deltas.dtype)

    boxes = boxes.astype(deltas.dtype, copy=False)#转换数据类型，使得二者一致

    #将anchor还原为（x,y,w,h）的格式
    widths = boxes[:, 2] - boxes[:, 0] + 1.0
    heights = boxes[:, 3] - boxes[:, 1] + 1.0
    ctr_x = boxes[:, 0] + 0.5 * widths
    ctr_y = boxes[:, 1] + 0.5 * heights
    #得到（x,y,w,h）方向上的偏移量
    dx = deltas[:, 0::4]
    dy = deltas[:, 1::4]
    dw = deltas[:, 2::4]
    dh = deltas[:, 3::4]

    pred_ctr_x = dx * widths[:, np.newaxis] + ctr_x[:, np.newaxis]#np.newaxis,表示将widths增加一维，使得其能够相加
    pred_ctr_y = dy * heights[:, np.newaxis] + ctr_y[:, np.newaxis]
    pred_w = np.exp(dw) * widths[:, np.newaxis]
    pred_h = np.exp(dh) * heights[:, np.newaxis]

    pred_boxes = np.zeros(deltas.shape, dtype=deltas.dtype)
    #最后返回的是左上和右下顶点的坐标[x1,y1,x2,y2]。
    # x1
    pred_boxes[:, 0::4] = pred_ctr_x - 0.5 * pred_w
    # y1
    pred_boxes[:, 1::4] = pred_ctr_y - 0.5 * pred_h
    # x2
    pred_boxes[:, 2::4] = pred_ctr_x + 0.5 * pred_w
    # y2
    pred_boxes[:, 3::4] = pred_ctr_y + 0.5 * pred_h

    return pred_boxes
# 将proposals的边界限制在图片内
# 调用格式 proposals = clip_boxes(proposals, im_info[:2])
def clip_boxes(boxes, im_shape):
    """
    Clip boxes to image boundaries.
    """

    # x1 >= 0
    boxes[:, 0::4] = np.maximum(np.minimum(boxes[:, 0::4], im_shape[1] - 1), 0)
    # y1 >= 0
    boxes[:, 1::4] = np.maximum(np.minimum(boxes[:, 1::4], im_shape[0] - 1), 0)
    # x2 < im_shape[1]
    boxes[:, 2::4] = np.maximum(np.minimum(boxes[:, 2::4], im_shape[1] - 1), 0)
    # y2 < im_shape[0]
    boxes[:, 3::4] = np.maximum(np.minimum(boxes[:, 3::4], im_shape[0] - 1), 0)
    return boxes