def __call__(self, loc, score, anchor, img_size, scale=1.):
if self.mode == "training":
n_pre_nms = self.n_train_pre_nms
n_post_nms = self.n_train_post_nms
else:
n_pre_nms = self.n_test_pre_nms
n_post_nms = self.n_test_post_nms
# 将RPN网络预测结果转化成建议框
roi = loc2bbox(anchor, loc)
# 利用slice进行分割，防止建议框超出图像边缘
roi[:, slice(0, 4, 2)] = np.clip(roi[:, slice(0, 4, 2)], 0,
img_size[1])
roi[:, slice(1, 4, 2)] = np.clip(roi[:, slice(1, 4, 2)], 0,
img_size[0])
# 宽高的最小值不可以小于16
min_size = self.min_size * scale
# 计算高宽
ws = roi[:, 2] - roi[:, 0]
hs = roi[:, 3] - roi[:, 1]
# 防止建议框过小
keep = np.where((hs >= min_size) & (ws >= min_size))[0]
roi = roi[keep, :]
score = score[keep]
# 取出成绩最好的一些建议框
order = score.ravel().argsort()[::-1]
if n_pre_nms > 0:
order = order[:n_pre_nms]
roi = roi[order, :]
roi = nms(roi, self.nms_thresh)
roi = torch.Tensor(roi)
roi = roi[:n_post_nms]
return roi
def __call__(self, loc, score, anchor, img_size, scale=1.):
"""
:param loc: 预测的偏移量 dx,dy,dw,dh (9*h*w,4)
:param score: 预测的正样本的概率（9*h*w,2)
:param anchor: anchorbox坐标（9*h*w,4)
:param img_size:
:param scale:
:return:
"""
if self.mode == "training":
n_pre_nms = self.n_train_pre_nms
n_post_nms = self.n_train_post_nms
else:
n_pre_nms = self.n_test_pre_nms
n_post_nms = self.n_test_post_nms
anchor = torch.from_numpy(anchor)
if loc.is_cuda:
anchor = anchor.cuda()
# -----------------------------------#
# 将RPN网络预测结果转化成建议框（x1,y1,x2,y2) anchor + loc = proposal 因为loc放的是偏移量
# -----------------------------------#
roi = loc2bbox(anchor, loc)
# -----------------------------------#
# 防止建议框超出图像边缘
# -----------------------------------#
roi[:, [0, 2]] = torch.clamp(roi[:, [0, 2]], min=0, max=img_size[1])
roi[:, [1, 3]] = torch.clamp(roi[:, [1, 3]], min=0, max=img_size[0])
# -----------------------------------#
# 建议框的宽高的最小值不可以小于16
# -----------------------------------#
min_size = self.min_size * scale
keep = torch.where(((roi[:, 2] - roi[:, 0]) >= min_size)
& ((roi[:, 3] - roi[:, 1]) >= min_size))[0]
roi = roi[keep, :]
score = score[keep]
# -----------------------------------#
# 根据得分进行排序，取出建议框
# -----------------------------------#
order = torch.argsort(score, descending=True)
if n_pre_nms > 0:
order = order[:n_pre_nms]
roi = roi[order, :]
score = score[order]
# -----------------------------------#
# 对建议框进行非极大抑制
# -----------------------------------#
keep = nms(roi, score, self.nms_thresh)
keep = keep[:n_post_nms] # 只取了300个proposal
roi = roi[keep]
return roi
def __call__(self, loc, score, anchor, img_size, scale=1.):
if self.mode == "training":
n_pre_nms = self.n_train_pre_nms
n_post_nms = self.n_train_post_nms
else:
n_pre_nms = self.n_test_pre_nms
n_post_nms = self.n_test_post_nms
anchor = torch.from_numpy(anchor)
if loc.is_cuda:
anchor = anchor.cuda()
#-----------------------------------#
# 将RPN网络预测结果转化成建议框
#-----------------------------------#
roi = loc2bbox(anchor, loc)
#-----------------------------------#
# 防止建议框超出图像边缘
#-----------------------------------#
roi[:, [0, 2]] = torch.clamp(roi[:, [0, 2]], min=0, max=img_size[1])
roi[:, [1, 3]] = torch.clamp(roi[:, [1, 3]], min=0, max=img_size[0])
#-----------------------------------#
# 建议框的宽高的最小值不可以小于16
#-----------------------------------#
min_size = self.min_size * scale
keep = torch.where(((roi[:, 2] - roi[:, 0]) >= min_size)
& ((roi[:, 3] - roi[:, 1]) >= min_size))[0]
roi = roi[keep, :]
score = score[keep]
#-----------------------------------#
# 根据得分进行排序，取出建议框
#-----------------------------------#
order = torch.argsort(score, descending=True)
if n_pre_nms > 0:
order = order[:n_pre_nms]
roi = roi[order, :]
score = score[order]
#-----------------------------------#
# 对建议框进行非极大抑制
#-----------------------------------#
keep = nms(roi, score, self.nms_thresh)
keep = keep[:n_post_nms]
roi = roi[keep]
return roi
def __call__(self, loc, score, anchor, img_size, scale=1.):
if self.mode == 'training':
n_pre_nms = self.n_train_pre_nms
n_post_nms = self.n_train_post_nms
else:
n_pre_nms = self.n_test_pre_nms
n_post_nms = self.n_test_post_nms
#将RPN网络预测结果转化为建议框
roi = loc2bbox(anchor, loc)
#利用slice进行分割，防止建议框超出图像边缘
roi[:, slice(0, 4, 2)] = np.clip(roi[:, slice(0, 4, 2)], 0,
img_size[1])
roi[:, slice(1, 4, 2)] = np.clip(roi[:, slice(1, 4, 2)], 0,
img_size[0])
#宽高的最小值不可以小于16
min_size = self.min_size * scale
#计算高宽
ws = roi[:, 2] - roi[:, 0]
hs = roi[:, 3] - roi[:, 1]
#防止建议框过小
keep = np.where((hs >= min_size) & (ws >= min_size))[0]
roi = roi[keep, :]
score = score[keep]
#取出成绩最好的一些建议框
oder = score.ravel().argsort()[::-1]
if n_pre_nms > 0:
oder = oder[:n_pre_nms]
roi = roi[oder, :]
score = score[oder]
keep = nms(
torch.from_numpy(roi).cuda(),
torch.from_numpy(score).cuda(), self.nms_thresh)
keep = keep[:n_post_nms]
roi = roi[keep.cpu().numpy()]
return roi