def to_Dets2(boxes, probs, img_ids, score_threshold=0.1):
"""for each box, there may be more than one class labels"""
boxes, probs, img_ids = everything2numpy([boxes, probs, img_ids])
Dets = []
for i in range(0, cfg.batch_size):
inds = np.where(img_ids == i)[0]
probs_ = probs[inds]
boxes_ = boxes[inds]
if probs_.shape[1] == 2:
cls_ids = np.ones((probs_.shape[0], ), dtype=np.int32)
cls_probs = probs_[:, 1]
dets = np.concatenate((boxes_.reshape(
-1, 4), cls_probs[:, np.newaxis], cls_ids[:, np.newaxis]),
axis=1)
else:
d0_inds, d1_inds = np.where(probs_[:, 1:] > score_threshold)
if d0_inds.size > 0:
cls_ids = d1_inds + 1
cls_probs = probs_[d0_inds, cls_ids]
boxes_ = boxes_[d0_inds, :]
dets = np.concatenate(
(boxes_.reshape(-1, 4), cls_probs[:, np.newaxis],
cls_ids[:, np.newaxis]),
axis=1)
else:
cls_ids = probs_[:, 1:].argmax(axis=1) + 1
cls_probs = probs_[np.arange(probs_.shape[0]), cls_ids]
dets = np.concatenate(
(boxes_.reshape(-1, 4), cls_probs[:, np.newaxis],
cls_ids[:, np.newaxis]),
axis=1)
Dets.append(dets)
return Dets
def to_Dets2_sigmoid(boxes, probs, img_ids, score_threshold=0.1):
boxes, probs, img_ids = everything2numpy([boxes, probs, img_ids])
Dets = []
for i in range(0, cfg.batch_size):
inds = np.where(img_ids == i)[0]
probs_ = probs[inds]
boxes_ = boxes[inds]
if probs_.ndim == 1 or probs_.shape[1] == 1:
cls_ids = np.ones((probs_.shape[0], ), dtype=np.int32)
cls_probs = probs_.view(-1)
dets = np.concatenate((boxes_.reshape(
-1, 4), cls_probs[:, np.newaxis], cls_ids[:, np.newaxis]),
axis=1)
else:
d0_inds, d1_inds = np.where(probs_ > score_threshold)
if d0_inds.size > 0:
cls_ids = d1_inds + 1
cls_probs = probs_[d0_inds, d1_inds]
boxes_ = boxes_[d0_inds, :]
dets = np.concatenate(
(boxes_.reshape(-1, 4), cls_probs[:, np.newaxis],
cls_ids[:, np.newaxis]),
axis=1)
else:
cls_ids = probs_.argmax(axis=1) + 1
cls_probs = probs_[np.arange(probs_.shape[0]), cls_ids - 1]
dets = np.concatenate(
(boxes_.reshape(-1, 4), cls_probs[:, np.newaxis],
cls_ids[:, np.newaxis]),
axis=1)
Dets.append(dets)
return Dets
def to_Dets_sigmoid(boxes, probs, img_ids):
"""
For each bbox, assign the class with the max prob.
NOTE: there is no background class,
so the implementation is slightly different.
"""
boxes, probs, img_ids = everything2numpy([boxes, probs, img_ids])
Dets = list()
for i in range(0, cfg.batch_size):
inds = np.where(img_ids == i)[0]
probs_ = probs[inds]
boxes_ = boxes[inds]
# !!! Difference is here. !!!
if probs_.ndim == 1 or probs_.shape[1] == 1:
cls_ids = np.ones((probs_.shape[0], ), dtype=np.int32)
cls_probs = probs_.view(-1)
else:
cls_ids = probs_.argmax(axis=1) + 1
cls_probs = probs_.max(axis=1)
dets = np.concatenate((boxes_.reshape(
-1, 4), cls_probs[:, np.newaxis], cls_ids[:, np.newaxis]),
axis=1)
Dets.append(dets)
# end_for
return Dets
def to_Dets(boxes, probs, img_ids):
"""
For each bbox, assign it with "the class" of the max prob.
"""
boxes, probs, img_ids = everything2numpy([boxes, probs, img_ids])
Dets = list()
for i in range(0, cfg.batch_size):
inds = np.where(img_ids == i)[0]
probs_ = probs[inds]
boxes_ = boxes[inds]
if probs_.shape[1] == 2:
cls_ids = np.ones((probs_.shape[0], ), dtype=np.int32)
cls_probs = probs_[:, 1]
else:
cls_ids = probs_[:, 1:].argmax(axis=1) + 1
cls_probs = probs_[np.arange(probs_.shape[0]), cls_ids]
dets = np.concatenate((boxes_.reshape(
-1, 4), cls_probs[:, np.newaxis], cls_ids[:, np.newaxis]),
axis=1)
Dets.append(dets)
# end_for
return Dets
## data loader
train_data = get_loader(cfg.data_dir, cfg.split, data_layer,
is_training=True, batch_size=cfg.batch_size, num_workers=cfg.data_workers)
ANCHORS = np.vstack([anc.reshape([-1, 4]) for anc in train_data.dataset.ANCHORS])
class_names = train_data.dataset.classes
print('dataset len: {}'.format(len(train_data.dataset)))
pixels = np.zeros((cfg.num_classes, ), dtype=np.int64)
instances = np.zeros((cfg.num_classes, ), dtype=np.int64)
timer = Timer()
timer.tic()
for step, batch in enumerate(train_data):
_, _, inst_masks, _, _, gt_boxes, _ = batch
inst_masks = \
everything2numpy(inst_masks)
for j, gt_box in enumerate(gt_boxes):
if gt_box.size > 0:
cls = gt_box[:, -1].astype(np.int32)
for i, c in enumerate(cls):
instances[c] += 1
m = inst_masks[j][i]
pixels[c] += m.sum()
t = timer.toc(False)
if step % 500 == 0:
print ('step: %d, instances: %d, pixels: %d, time: %.2fs' % (step, instances.sum(), pixels.sum(), t))
with open("statistics", "wb") as f:
pickle.dump({
writer.add_summary(s, float(global_step))
# Save Model
if step % 5000 == 0 and global_step != 0:
if not cfg.save_prefix:
save_path = os.path.join()
else:
save_path = os.path.join()
save_net()
print('')
# Draw Detection Results (Stage-1, Stage-2)
if global_step % cfg.log_image == 0:
summary_out = []
input_np = everything2numpy(input)
# Get Detection Results
dets_dict = model_ori.get_final_results()
for key, dets in dets_dict.iteritems():
Is = single_shot.draw_detection()
Is = Is.astype(np.uint8)
summary_out += log_images()
# Draw Ground-Truth
Is = single_shot.draw_gtboxes()
Is = Is.astype(np.uint8)
summary_out += log_images()
summary = model_ori.get_summaries()
for s in summary:
def main():
# config model and lr
num_anchors = len(cfg.anchor_ratios) * len(cfg.anchor_scales[0]) * len(cfg.anchor_shift) \
if isinstance(cfg.anchor_scales[0], list) else \
len(cfg.anchor_ratios) * len(cfg.anchor_scales)
resnet = resnet50 if cfg.backbone == 'resnet50' else resnet101
detection_model = MaskRCNN if cfg.model_type.lower(
) == 'maskrcnn' else RetinaNet
model = detection_model(resnet(pretrained=True, maxpool5=cfg.maxpool5),
num_classes=cfg.num_classes,
num_anchors=num_anchors,
strides=cfg.strides,
in_channels=cfg.in_channels,
f_keys=cfg.f_keys,
num_channels=256,
is_training=False,
activation=cfg.class_activation)
lr = cfg.lr
start_epoch = 0
if cfg.restore is not None:
meta = load_net(cfg.restore, model)
print(meta)
if meta[0] >= 0:
start_epoch = meta[0] + 1
lr = meta[1]
print('Restored from %s, starting from %d epoch, lr:%.6f' %
(cfg.restore, start_epoch, lr))
else:
raise ValueError('restore is not set')
model.cuda()
model.eval()
class_names = test_data.dataset.classes
print('dataset len: {}'.format(len(test_data.dataset)))
tb_dir = os.path.join(cfg.train_dir, cfg.backbone + '_' + cfg.datasetname,
'test', time.strftime("%h%d_%H"))
writer = tbx.FileWriter(tb_dir)
# main loop
timer_all = Timer()
timer_post = Timer()
all_results1 = []
all_results2 = []
all_results_gt = []
for step, batch in enumerate(test_data):
timer_all.tic()
# NOTE: Targets is in NHWC order!!
# input, anchors_np, im_scale_list, image_ids, gt_boxes_list = batch
# input = everything2cuda(input)
input_t, anchors_np, im_scale_list, image_ids, gt_boxes_list = batch
input = everything2cuda(input_t, volatile=True)
outs = model(input, gt_boxes_list=None, anchors_np=anchors_np)
if cfg.model_type == 'maskrcnn':
rpn_logit, rpn_box, rpn_prob, rpn_labels, rpn_bbtargets, rpn_bbwghts, anchors, \
rois, roi_img_ids, rcnn_logit, rcnn_box, rcnn_prob, rcnn_labels, rcnn_bbtargets, rcnn_bbwghts = outs
outputs = [
rois, roi_img_ids, rpn_logit, rpn_box, rpn_prob, rcnn_logit,
rcnn_box, rcnn_prob, anchors
]
targets = []
elif cfg.model_type == 'retinanet':
rpn_logit, rpn_box, rpn_prob, _, _, _ = outs
outputs = [rpn_logit, rpn_box, rpn_prob]
else:
raise ValueError('Unknown model type: %s' % cfg.model_type)
timer_post.tic()
dets_dict = model.get_final_results(
outputs,
everything2cuda(anchors_np),
score_threshold=0.01,
max_dets=cfg.max_det_num * cfg.batch_size,
overlap_threshold=cfg.overlap_threshold)
if 'stage1' in dets_dict:
Dets = dets_dict['stage1']
else:
raise ValueError('No stage1 results:', dets_dict.keys())
Dets2 = dets_dict['stage2'] if 'stage2' in dets_dict else Dets
t3 = timer_post.toc()
t = timer_all.toc()
formal_res1 = dataset.to_detection_format(copy.deepcopy(Dets),
image_ids, im_scale_list)
formal_res2 = dataset.to_detection_format(copy.deepcopy(Dets2),
image_ids, im_scale_list)
all_results1 += formal_res1
all_results2 += formal_res2
Dets_gt = []
for gb in gt_boxes_list:
cpy_mask = gb[:, 4] >= 1
gb = gb[cpy_mask]
n = cpy_mask.astype(np.int32).sum()
res_gt = np.zeros((n, 6))
res_gt[:, :4] = gb[:, :4]
res_gt[:, 4] = 1.
res_gt[:, 5] = gb[:, 4]
Dets_gt.append(res_gt)
formal_res_gt = dataset.to_detection_format(Dets_gt, image_ids,
im_scale_list)
all_results_gt += formal_res_gt
if step % cfg.log_image == 0:
input_np = everything2numpy(input)
summary_out = []
Is = single_shot.draw_detection(input_np,
Dets,
class_names=class_names)
Is = Is.astype(np.uint8)
summary_out += log_images(Is, image_ids, step, prefix='Detection/')
Is = single_shot.draw_detection(input_np,
Dets2,
class_names=class_names)
Is = Is.astype(np.uint8)
summary_out += log_images(Is,
image_ids,
step,
prefix='Detection2/')
Imgs = single_shot.draw_gtboxes(input_np,
gt_boxes_list,
class_names=class_names)
Imgs = Imgs.astype(np.uint8)
summary_out += log_images(Imgs,
image_ids,
float(step),
prefix='GT')
for s in summary_out:
writer.add_summary(s, float(step))
if step % cfg.display == 0:
print(time.strftime("%H:%M:%S ") +
'Epoch %d iter %d: speed %.3fs (%.3fs)' % (0, step, t, t3) +
' ImageIds: ' + ', '.join(str(s) for s in image_ids),
end='\r')
res_dict = {
'stage1': all_results1,
'stage2': all_results2,
'gt': all_results_gt
}
return res_dict
def main():
# config model and lr
num_anchors = len(cfg.anchor_ratios) * len(cfg.anchor_scales[0]) \
if isinstance(cfg.anchor_scales[0], list) else \
len(cfg.anchor_ratios) * len(cfg.anchor_scales)
resnet = resnet50 if cfg.backbone == 'resnet50' else resnet101
detection_model = MaskRCNN if cfg.model_type.lower(
) == 'maskrcnn' else RetinaNet
model = detection_model(resnet(pretrained=True),
num_classes=cfg.num_classes,
num_anchors=num_anchors,
strides=cfg.strides,
in_channels=cfg.in_channels,
f_keys=cfg.f_keys,
num_channels=256,
is_training=False,
activation=cfg.class_activation)
lr = cfg.lr
start_epoch = 0
if cfg.restore is not None:
meta = load_net(cfg.restore, model)
print(meta)
if meta[0] >= 0:
start_epoch = meta[0] + 1
lr = meta[1]
print('Restored from %s, starting from %d epoch, lr:%.6f' %
(cfg.restore, start_epoch, lr))
else:
raise ValueError('restore is not set')
model.cuda()
model.eval()
ANCHORS = np.vstack(
[anc.reshape([-1, 4]) for anc in test_data.dataset.ANCHORS])
model.anchors = everything2cuda(ANCHORS.astype(np.float32))
class_names = test_data.dataset.classes
print('dataset len: {}'.format(len(test_data.dataset)))
tb_dir = os.path.join(cfg.train_dir, cfg.backbone + '_' + cfg.datasetname,
'test', time.strftime("%h%d_%H"))
writer = tbx.FileWriter(tb_dir)
summary_out = []
# main loop
timer_all = Timer()
timer_post = Timer()
all_results1 = []
all_results2 = []
all_results_gt = []
for step, batch in enumerate(test_data):
timer_all.tic()
# NOTE: Targets is in NHWC order!!
input, image_ids, gt_boxes_list, image_ori = batch
input = everything2cuda(input)
outs = model(input)
timer_post.tic()
dets_dict = model.get_final_results(
score_threshold=0.05,
max_dets=cfg.max_det_num * cfg.batch_size,
overlap_threshold=cfg.overlap_threshold)
if 'stage1' in dets_dict:
Dets = dets_dict['stage1']
else:
raise ValueError('No stage1 results:', dets_dict.keys())
Dets2 = dets_dict['stage2'] if 'stage2' in dets_dict else Dets
t3 = timer_post.toc()
t = timer_all.toc()
formal_res1 = dataset.to_detection_format(
copy.deepcopy(Dets),
image_ids,
ori_sizes=[im.shape for im in image_ori])
formal_res2 = dataset.to_detection_format(
copy.deepcopy(Dets2),
image_ids,
ori_sizes=[im.shape for im in image_ori])
all_results1 += formal_res1
all_results2 += formal_res2
if step % cfg.log_image == 0:
input_np = everything2numpy(input)
summary_out = []
Is = single_shot.draw_detection(input_np,
Dets,
class_names=class_names)
Is = Is.astype(np.uint8)
summary_out += log_images(Is, image_ids, step, prefix='Detection/')
Is = single_shot.draw_detection(input_np,
Dets2,
class_names=class_names)
Is = Is.astype(np.uint8)
summary_out += log_images(Is,
image_ids,
step,
prefix='Detection2/')
Imgs = single_shot.draw_gtboxes(input_np,
gt_boxes_list,
class_names=class_names)
Imgs = Imgs.astype(np.uint8)
summary_out += log_images(Imgs,
image_ids,
float(step),
prefix='GT')
for s in summary_out:
writer.add_summary(s, float(step))
if step % cfg.display == 0:
print(time.strftime("%H:%M:%S ") +
'Epoch %d iter %d: speed %.3fs (%.3fs)' % (0, step, t, t3) +
' ImageIds: ' + ', '.join(str(s) for s in image_ids),
end='\r')
res_dict = {
'stage1': all_results1,
'stage2': all_results2,
'gt': all_results_gt
}
return res_dict