# --------------------------------------------------------
# Fast R-CNN
# Copyright (c) 2015 Microsoft
# Licensed under The MIT License [see LICENSE for details]
# Written by Sergey Karayev
# --------------------------------------------------------
cimport cython
import numpy as np
cimport numpy as np
DTYPE = np.float
ctypedef np.float_t DTYPE_t
# rpn:region proposal network
# IoU:Intersection-over-Union,交集并集之比
# GT:ground truth,GT boxes (x1, y1, x2, y2, label),左上的坐标 和 右下的坐标 + label
# im:image
# ROI:region of interest
# bbox:bounding-box
# regression:和one hot的classification其实区别不大,只是target是(可能为0和1以及)其他数字
# anchor:一个box,通过从上一个卷积层的结果 滑动得到很多个,每个要和 GT的box 计算bbox_overlap
# bbox_overlap:貌似就是IoU,源码如下
def bbox_overlaps(
np.ndarray[DTYPE_t, ndim=2] boxes,
np.ndarray[DTYPE_t, ndim=2] query_boxes):
"""
Parameters
----------
boxes: (N, 4) ndarray of float
query_boxes: (K, 4) ndarray of float
Returns
-------
overlaps: (N, K) ndarray of overlap between boxes and query_boxes
"""
#记录anchor的个数
cdef unsigned int N = boxes.shape[0]
#记录ground-truth的个数
cdef unsigned int K = query_boxes.shape[0]
#创建一个全零arrayoverlaps
cdef np.ndarray[DTYPE_t, ndim=2] overlaps = np.zeros((N, K), dtype=DTYPE)
cdef DTYPE_t iw, ih, box_area
cdef DTYPE_t ua
cdef unsigned int k, n
for k in range(K):
box_area = (#query_boxes(GT)的面积
(query_boxes[k, 2] - query_boxes[k, 0] + 1) *#列数(x)
(query_boxes[k, 3] - query_boxes[k, 1] + 1)#行数
)
#每一个anchor与当前GT比一下
for n in range(N):
iw = (#计算重叠部分的宽
min(boxes[n, 2], query_boxes[k, 2]) -
max(boxes[n, 0], query_boxes[k, 0]) + 1
)
if iw > 0:
ih = (#计算重叠部分的高
min(boxes[n, 3], query_boxes[k, 3]) -
max(boxes[n, 1], query_boxes[k, 1]) + 1
)
if ih > 0:
ua = float(#计算(anchor+GT-重叠)的面积
(boxes[n, 2] - boxes[n, 0] + 1) *
(boxes[n, 3] - boxes[n, 1] + 1) +
box_area - iw * ih
)
#对应(n,k)元素存入(重叠/(anchor+GT-重叠))面积
overlaps[n, k] = iw * ih / ua
return overlaps
Faster-RCNN_TF代码解读19:bbox.pyx
最新推荐文章于 2024-03-04 20:22:56 发布