def __init__(self, anchor_num=5, feature_in=256, feature_out=256):
super(UP, self).__init__()
self.anchor_num = anchor_num
self.feature_in = feature_in
self.feature_out = feature_out
self.cls_output = 2 * self.anchor_num
self.loc_output = 4 * self.anchor_num
self.cls = DepthCorr(feature_in, feature_out, self.cls_output)
self.loc = DepthCorr(feature_in, feature_out, self.loc_output)
def __init__(self, anchor_num=5, feature_in=256, feature_out=256):
super(UP, self).__init__()
self.anchor_num = anchor_num
self.feature_in = feature_in
self.feature_out = feature_out
#分类输出，正类还是负类，故*2；定位anchor需要两对坐标四个输出，故*3
self.cls_output = 2 * self.anchor_num
self.loc_output = 4 * self.anchor_num
self.cls = DepthCorr(feature_in, feature_out, self.cls_output)
self.loc = DepthCorr(feature_in, feature_out, self.loc_output)
def __init__(self, anchor_num=5, feature_in=256, feature_out=256):
super(UP, self).__init__()
self.anchor_num = anchor_num
self.feature_in = feature_in
self.feature_out = feature_out
self.cls_output = 2 * self.anchor_num
self.loc_output = 4 * self.anchor_num
# 下面两行本质上是实例化了“两个”分支，分别是rpn的分类分支和回归分支
self.cls = DepthCorr(
feature_in, feature_out, self.cls_output
) # 这里的feature_in, feature_out, self.cls_output是指输入输出的通道数
self.loc = DepthCorr(feature_in, feature_out, self.loc_output)
def __init__(self):
super(KpCorr, self).__init__()
self.kp = DepthCorr(256, 512, 1024)
def __init__(self, oSz=63):
super(MaskCorr, self).__init__()
self.oSz = oSz
self.mask = DepthCorr(256, 256, self.oSz**2 * 17)
def __init__(self, oSz=63):
super(MaskCorr, self).__init__()
self.oSz = oSz
self.kp_direct_reg = DepthCorr(256, 256, 17*2)
self.kp_heatmap = DepthCorr(256, 256, oSz**2)
def __init__(self, oSz=63): # 为什么是63*63？
super(MaskCorr, self).__init__()
self.oSz = oSz
self.mask = DepthCorr(
256, 256, self.oSz**
2) # DepthCorr类包含了从调整层adjust到depth-wise-corr再到头部结构hΦ/bσ/sφ
def __init__(self):
super(PoseCorr, self).__init__()
self.kp = DepthCorr(256, 512, 512)