#功能描述:生成多尺度、多宽高比的anchors。 # 尺度为:128,256,512; 宽高比为:1:2,1:1,2:1 import numpy as np #提供矩阵运算功能的库 #生成anchors总函数:ratios为一个列表,表示宽高比为:1:2,1:1,2:1 #2**x表示:2^x,scales:[2^3 2^4 2^5],即:[8 16 32] def generate_anchors(base_size=16, ratios=[0.5, 1, 2], scales=2**np.arange(3, 6)): """ Generate anchor (reference) windows by enumerating aspect ratios X scales wrt a reference (0, 0, 15, 15) window. """ base_anchor = np.array([1, 1, base_size, base_size]) - 1 #新建一个数组:base_anchor:[0 0 15 15] ratio_anchors = _ratio_enum(base_anchor, ratios) #枚举各种宽高比 anchors = np.vstack([_scale_enum(ratio_anchors[i, :], scales) #枚举各种尺度,vstack:竖向合并数组 for i in range(ratio_anchors.shape[0])]) #shape[0]:读取矩阵第一维长度,其值为3 return anchors #用于返回width,height,(x,y)中心坐标(对于一个anchor窗口) def _whctrs(anchor): """ Return width, height, x center, and y center for an anchor (window). """ #anchor:存储了窗口左上角,右下角的坐标 w = anchor[2] - anchor[0] + 1 h = anchor[3] - anchor[1] + 1 x_ctr = anchor[0] + 0.5 * (w - 1) #anchor中心点坐标 y_ctr = anchor[1] + 0.5 * (h - 1) return w, h, x_ctr, y_ctr #给定一组宽高向量,输出各个anchor,即预测窗口,**输出anchor的面积相等,只是宽高比不同** def _mkanchors(ws, hs, x_ctr, y_ctr): #ws:[23 16 11],hs:[12 16 22],ws和hs一一对应。 """ Given a vector of widths (ws) and heights (hs) around a center (x_ctr, y_ctr), output a set of anchors (windows). """ ws = ws[:, np.newaxis] #newaxis:将数组转置 hs = hs[:, np.newaxis] anchors = np.hstack((x_ctr - 0.5 * (ws - 1), #hstack、vstack:合并数组 y_ctr - 0.5 * (hs - 1), #anchor:[[-3.5 2 18.5 13] x_ctr + 0.5 * (ws - 1), # [0 0 15 15] y_ctr + 0.5 * (hs - 1))) # [2.5 -3 12.5 18]] return anchors #枚举一个anchor的各种宽高比,anchor[0 0 15 15],ratios[0.5,1,2] def _ratio_enum(anchor, ratios): """ 列举关于一个anchor的三种宽高比 1:2,1:1,2:1 Enumerate a set of anchors for each aspect ratio wrt an anchor. """ w, h, x_ctr, y_ctr = _whctrs(anchor) #返回宽高和中心坐标,w:16,h:16,x_ctr:7.5,y_ctr:7.5 size = w * h #size:16*16=256 size_ratios = size / ratios #256/ratios[0.5,1,2]=[512,256,128] #round()方法返回x的四舍五入的数字,sqrt()方法返回数字x的平方根 ws = np.round(np.sqrt(size_ratios)) #ws:[23 16 11] hs = np.round(ws * ratios) #hs:[12 16 22],ws和hs一一对应。as:23&12 anchors = _mkanchors(ws, hs, x_ctr, y_ctr) #给定一组宽高向量,输出各个预测窗口 return anchors #枚举一个anchor的各种尺度,以anchor[0 0 15 15]为例,scales[8 16 32] def _scale_enum(anchor, scales): """ 列举关于一个anchor的三种尺度 128*128,256*256,512*512 Enumerate a set of anchors for each scale wrt an anchor. """ w, h, x_ctr, y_ctr = _whctrs(anchor) #返回宽高和中心坐标,w:16,h:16,x_ctr:7.5,y_ctr:7.5 ws = w * scales #[128 256 512] hs = h * scales #[128 256 512] anchors = _mkanchors(ws, hs, x_ctr, y_ctr) #[[-56 -56 71 71] [-120 -120 135 135] [-248 -248 263 263]] return anchors if __name__ == '__main__': #主函数 import time t = time.time() a = generate_anchors() #生成anchor(窗口) print(time.time() - t ) #显示时间 print(a) from IPython import embed; embed()
普及一下生成anchor的函数实现原理,追随源码(generate_anchors.py):
def generate_anchors(base_size=16, ratios=[0.5, 1, 2],
scales=2**np.arange(3, 6)):
base_anchor = np.array([1, 1, base_size, base_size]) - 1
ratio_anchors = _ratio_enum(base_anchor, ratios)
anchors = np.vstack([_scale_enum(ratio_anchors[i, :], scales)
for i in xrange(ratio_anchors.shape[0])])
return anchors
这个函数就是生成九个anchors的函数,首先有一个base_anchor坐标为[0,0,15,15],因为电脑是从0开始计数的,其实是[1,1,16,16],先调用_ratio_enum
def _ratio_enum(anchor, ratios):
w, h, x_ctr, y_ctr = _whctrs(anchor)
size = w * h
size_ratios = size / ratios
ws = np.round(np.sqrt(size_ratios))
hs = np.round(ws * ratios)
anchors = _mkanchors(ws, hs, x_ctr, y_ctr)
return anchors
在这个函数里先调用了_whctrs,作用是得到anchor 的四个参数,宽度w=16,高度h=16,中心点坐标x=7.5,y=7.5,之后做了一系列数学计算,最终结果为ws=[23,16,11], hs=[12,16,22],调用_mkanchors
def _mkanchors(ws, hs, x_ctr, y_ctr):
ws = ws[:, np.newaxis]
hs = hs[:, np.newaxis]
anchors = np.hstack((x_ctr - 0.5 * (ws - 1),
y_ctr - 0.5 * (hs - 1),
x_ctr + 0.5 * (ws - 1),
y_ctr + 0.5 * (hs - 1)))
return anchors
在这个函数的里面有个np.newaxis,是增加数据的维数的意思,原来ws,hs均为一组数据(只有一个维度),之后变成一个3行,1列(虽然是1,但是也是个维度)的二维数据,即变成了ws={[23],[16],[11]}, hs={[12],[16],[22]},最后一个函数不想细说,总之就是变成了[■(-3.5&2&18.5@[email protected]&-3&12.5) ■(13@15@18)],即ratio_anchors,我们在回到最基本的那个函数,接下来又调用了一个函数
anchors = np.vstack([_scale_enum(ratio_anchors[i, :], scales)
for i in xrange(ratio_anchors.shape[0])])
在这里ratio_anchors.shape[0]指的是3行4列的3,也就是一行一行的输送给_scale_enum函数
def _scale_enum(anchor, scales):
w, h, x_ctr, y_ctr = _whctrs(anchor)
ws = w * scales
hs = h * scales
anchors = _mkanchors(ws, hs, x_ctr, y_ctr)
return anchors
由最开始的那个函数的参数可知,scales为2的3,4,5次方即[8,16,32]。首先,先得到四个参数(w, h, x_ctr, y_ctr)=[■(23&12&7.5@16&16&7.5@11&22&7.5) ■([email protected]@7.5)]所以ws=[184,368,736], [128,256,512],[88,176,352], hs=[96,192,384], [128,256,512], [176,352,704], 这是9组对应的宽和高,其实每次只能得到三组,我是直接把循环三次的结果写了出来,将这9组的数据送到_mkanchors之后得到9个anchors,分别为:
-84.0 -40.0 99 55
-176.0 -88.0 191 103
-360.0 -184.0 375 199
-56.0 -56.0 71 71
-120.0 -120.0 135 135
-248.0 -248.0 263 263
-36.0 -80.0 51 95
-80.0 -168.0 95 183
-168.0 -344.0 183 359
转换成我们需要的四个参数分别为:
ratio = 0.5
(184.0, 96.0, 7.5, 7.5)
(368.0, 192.0, 7.5, 7.5)
(736.0, 384.0, 7.5, 7.5)
ratio = 1.0
(128.0, 128.0, 7.5, 7.5)
(256.0, 256.0, 7.5, 7.5)
(512.0, 512.0, 7.5, 7.5)
ratio = 2.0
(88.0, 176.0, 7.5, 7.5)
(176.0, 352.0, 7.5, 7.5)
(352.0, 704.0, 7.5, 7.5)