def draw_final_outputs_blackwhite(img, results):
"""
Args:
results: [DetectionResult]
"""
img_bw = img.mean(axis=2)
img_bw = np.stack([img_bw] * 3, axis=2)
if len(results) == 0:
return img_bw
boxes = np.asarray([r.box for r in results])
all_masks = [r.mask for r in results]
if all_masks[0] is not None:
m = all_masks[0] > 0
for m2 in all_masks[1:]:
m = m | (m2 > 0)
img_bw[m] = img[m]
tags = [
"{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id], r.score)
for r in results
]
ret = viz.draw_boxes(img_bw, boxes, tags)
return ret
def draw_final_outputs(img, results):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
# Display in largest to smallest order to reduce occlusion
boxes = np.asarray([r.box for r in results])
areas = np_area(boxes)
sorted_inds = np.argsort(-areas)
ret = img
tags = []
for result_id in sorted_inds:
r = results[result_id]
if r.mask is not None:
ret = draw_mask(ret, r.mask)
for r in results:
tags.append("{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id],
r.score))
ret = viz.draw_boxes(ret, boxes, tags)
return ret
def draw_final_outputs(img, results,objectfile):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
tags = []
for r in results:
tags.append(
"{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id], r.score))
objectfile.write(cfg.DATA.CLASS_NAMES[r.class_id])
objectfile.write('\t')
objectfile.write('-1')
objectfile.write('\t')
objectfile.write('-1')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('{:d}'.format(int(r[0][0])))
objectfile.write('\t')
objectfile.write('{:d}'.format(int(r[0][1])))
objectfile.write('\t')
objectfile.write('{:d}'.format(int(r[0][2])))
objectfile.write('\t')
objectfile.write('{:d}'.format(int(r[0][3])))
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('0.0')
objectfile.write('\t')
objectfile.write('{:.2f}'.format(r.score))
objectfile.write('\t')
objectfile.write('\n')
boxes = np.asarray([r.box for r in results])
ret = viz.draw_boxes(img, boxes, tags)
#ret = viz.draw_boxes(img, boxes)
objectfile.close()
for r in results:
if r.mask is not None:
ret = draw_mask(ret, r.mask)
return ret
def draw_predictions(img, boxes, scores):
"""
Args:
boxes: kx4
scores: kxC
"""
if len(boxes) == 0:
return img
labels = scores.argmax(axis=1)
scores = scores.max(axis=1)
tags = ["{},{:.2f}".format(cfg.DATA.CLASS_NAMES[lb], score) for lb, score in zip(labels, scores)]
return viz.draw_boxes(img, boxes, tags)
def draw_predictions(img, boxes, scores):
"""
Args:
boxes: kx4
scores: kxC
"""
if len(boxes) == 0:
return img
labels = scores.argmax(axis=1)
scores = scores.max(axis=1)
tags = ["{},{:.2f}".format(cfg.DATA.CLASS_NAMES[lb], score) for lb, score in zip(labels, scores)]
return viz.draw_boxes(img, boxes, tags)
def draw_predictions(img, boxes, scores):
"""
Args:
boxes: kx4
scores: kxC
"""
if len(boxes) == 0:
return img
labels = scores.argmax(axis=1)
scores = scores.max(axis=1)
tags = ["{},{:.2f}".format(COCOMeta.class_names[lb], score) for lb, score in zip(labels, scores)]
return viz.draw_boxes(img, boxes, tags)
def __getitem__(self, imageid):
filename = self.cocodb.lookup[imageid]['file_name']
I = cv2.imread(self.prefix + '/' + filename)
obj = self._d[imageid]
labels = ['%d:%.2f' % (ll, ss) for ll, ss in zip(obj.label, obj.score)]
print(obj.bbox)
boxes = [[bb[0], bb[1], bb[0] + bb[2], bb[1] + bb[3]]
for bb in obj.bbox]
I = viz.draw_boxes(I, boxes, labels)
print('total # boxes:%d' % len(boxes))
return I
def draw_final_outputs(img, results):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
tags = []
for label, _, score in results:
tags.append(
"{},{:.2f}".format(config.CLASS_NAMES[label], score))
boxes = np.asarray([x.box for x in results])
return viz.draw_boxes(img, boxes, tags)
def draw_annotation(img, boxes, klass, is_crowd=None):
labels = []
assert len(boxes) == len(klass)
if is_crowd is not None:
assert len(boxes) == len(is_crowd)
for cls, crd in zip(klass, is_crowd):
clsname = config.CLASS_NAMES[cls]
if crd == 1:
clsname += ';Crowd'
labels.append(clsname)
else:
for cls in klass:
labels.append(config.CLASS_NAMES[cls])
img = viz.draw_boxes(img, boxes, labels)
return img
def draw_final_outputs(img, results):
"""
Args:
results: [DetectionResult]
"""
all_boxes = []
all_tags = []
for class_id, boxes, scores in results:
all_boxes.extend(boxes)
all_tags.extend(
["{},{:.2f}".format(COCOMeta.class_names[class_id], sc) for sc in scores])
all_boxes = np.asarray(all_boxes)
if all_boxes.shape[0] == 0:
return img
return viz.draw_boxes(img, all_boxes, all_tags)
def draw_annotation(img, boxes, klass, is_crowd=None):
labels = []
assert len(boxes) == len(klass)
if is_crowd is not None:
assert len(boxes) == len(is_crowd)
for cls, crd in zip(klass, is_crowd):
clsname = config.CLASS_NAMES[cls]
if crd == 1:
clsname += ';Crowd'
labels.append(clsname)
else:
for cls in klass:
labels.append(config.CLASS_NAMES[cls])
img = viz.draw_boxes(img, boxes, labels)
return img
def draw_mask(im, mask, box, label, alpha=0.5, color=None):
"""
Overlay a mask on top of the image.
Args:
im: a 3-channel uint8 image in BGR
mask: a binary 1-channel image of the same size
color: if None, will choose automatically
"""
if color is None:
color = PALETTE_RGB[np.random.choice(len(PALETTE_RGB))][::-1]
im = np.where(np.repeat((mask > 0)[:, :, None], 3, axis=2),
im * (1 - alpha) + color * alpha, im)
im = im.astype('uint8')
color_tuple = tuple([int(c) for c in color])
im = viz.draw_boxes(im, box[np.newaxis, :], [label], color=color_tuple)
return im
def draw_annotation(img, boxes, klass, is_crowd=None):
"""Will not modify img"""
labels = []
assert len(boxes) == len(klass)
if is_crowd is not None:
assert len(boxes) == len(is_crowd)
for cls, crd in zip(klass, is_crowd):
clsname = cfg.DATA.CLASS_NAMES[cls]
if crd == 1:
clsname += ";Crowd"
labels.append(clsname)
else:
for cls in klass:
labels.append(cfg.DATA.CLASS_NAMES[cls])
img = viz.draw_boxes(img, boxes, labels)
return img
def draw_proposal_recall(img, proposals, proposal_scores, gt_boxes):
"""
Draw top3 proposals for each gt.
Args:
proposals: NPx4
proposal_scores: NP
gt_boxes: NG
"""
box_ious = np_iou(gt_boxes, proposals) # ng x np
box_ious_argsort = np.argsort(-box_ious, axis=1)
good_proposals_ind = box_ious_argsort[:, :3] # for each gt, find 3 best proposals
good_proposals_ind = np.unique(good_proposals_ind.ravel())
proposals = proposals[good_proposals_ind, :]
tags = list(map(str, proposal_scores[good_proposals_ind]))
img = viz.draw_boxes(img, proposals, tags)
return img, good_proposals_ind
def draw_proposal_recall(img, proposals, proposal_scores, gt_boxes):
"""
Draw top3 proposals for each gt.
Args:
proposals: NPx4
proposal_scores: NP
gt_boxes: NG
"""
box_ious = np_iou(gt_boxes, proposals) # ng x np
box_ious_argsort = np.argsort(-box_ious, axis=1)
good_proposals_ind = box_ious_argsort[:, :3] # for each gt, find 3 best proposals
good_proposals_ind = np.unique(good_proposals_ind.ravel())
proposals = proposals[good_proposals_ind, :]
tags = list(map(str, proposal_scores[good_proposals_ind]))
img = viz.draw_boxes(img, proposals, tags)
return img, good_proposals_ind
def draw_final_outputs(img, results):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
tags = []
for r in results:
tags.append(
"{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id], r.score))
boxes = np.asarray([r.box for r in results])
ret = viz.draw_boxes(img, boxes, tags)
for r in results:
if r.mask is not None:
ret = draw_mask(ret, r.mask)
return ret
def draw_final_outputs(img, results):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
tags = []
for r in results:
tags.append("{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id],
r.score))
boxes = np.asarray([r.box for r in results])
ret = viz.draw_boxes(img, boxes) #, tags)
for r in results:
if r.mask is not None:
ret = draw_mask(ret, r.mask)
return ret
def draw_annotation(img, boxes, klass, polygons=None, is_crowd=None):
"""Will not modify img"""
labels = []
assert len(boxes) == len(klass)
if is_crowd is not None:
assert len(boxes) == len(is_crowd)
for cls, crd in zip(klass, is_crowd):
clsname = cfg.DATA.CLASS_NAMES[cls]
if crd == 1:
clsname += ';Crowd'
labels.append(clsname)
else:
for cls in klass:
labels.append(cfg.DATA.CLASS_NAMES[cls])
img = viz.draw_boxes(img, boxes, labels)
if polygons is not None:
for p in polygons:
mask = polygons_to_mask(p, img.shape[0], img.shape[1])
img = draw_mask(img, mask)
return img
def draw_final_outputs(img, results, tags_on=True, bb_list_input=False):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
if tags_on:
tags = []
for r in results:
tags.append("{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id],
r.score))
else:
tags = None
if bb_list_input:
boxes = np.asarray(results)
else:
boxes = np.asarray([r.box for r in results])
ret = viz.draw_boxes(img, boxes, tags)
return ret
def draw_outputs(img, final_boxes, final_scores, final_labels, threshold=0.8):
"""
Args:
results: [DetectionResult]
"""
results = [DetectionResult(*args) for args in
zip(final_boxes, final_scores, final_labels,
[None] * len(final_labels)) if args[1]>threshold]
if len(results) == 0:
return img
tags = []
for r in results:
tags.append(
"{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id], r.score))
boxes = np.asarray([r.box for r in results])
ret = viz.draw_boxes(img, boxes, tags)
for r in results:
if r.mask is not None:
ret = draw_mask(ret, r.mask)
return ret
def draw_final_outputs(img, results, show_ids=None):
"""
Args:
results: [DetectionResult]
"""
if len(results) == 0:
return img
if show_ids is not None:
assert isinstance(show_ids, set)
results = [r for r in results if r.class_id in show_ids]
tags = []
for r in results:
tags.append(
"{},{:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id], r.score))
boxes = np.asarray([r.box for r in results])
ret = viz.draw_boxes(img, boxes, tags)
# for r in results:
# if r.mask is not None:
# ret = draw_mask(ret, r.mask)
return ret
except StopIteration:
break
orig_shape = batch_image.shape[:2]
feed_dict = {image_P: batch_image}
final_boxes_, final_labels_, final_probs_ = sess.run(
[final_boxes, final_labels, final_probs], feed_dict)
final_boxes_ = clip_boxes(final_boxes_, orig_shape)
final_boxes_ = sess.run(final_boxes_)
final_boxes_ = final_boxes_.astype('int32')
if np.any(final_boxes_):
tags = [
"{},{:.2f}".format(cfg.DATA.CLASS_NAMES[lb], score)
for lb, score in zip(final_labels_, final_probs_)
]
final_viz = viz.draw_boxes(batch_image, final_boxes_, tags)
gt_viz = draw_annotation(batch_image, batch_gt_boxes, batch_gt_labels)
img_out = np.hstack((final_viz, gt_viz))
imageio.imwrite(os.path.join(save_path, str(iter) + ".jpg"), img_out)
Detection = []
for ik in range(final_boxes_.shape[0]):
Detection.append([
cfg.DATA.CLASS_NAMES[final_labels_[ik]],
float(final_probs_[ik]), final_boxes_[ik, 0], final_boxes_[ik, 1],
final_boxes_[ik, 2] - final_boxes_[ik, 0],
final_boxes_[ik, 3] - final_boxes_[ik, 1]
])
Detection = np.array(Detection)
#np.savetxt(os.path.join(save_folder, 'detections/',str(iter) + '.txt'), Detection, fmt='%s %1.2f %1.0f %1.0f %1.0f %1.0f')
def visualize_dataflow2(cfg, unlabled2017_used=True, VISPATH="./", maxvis=50):
"""Visualize the dataflow with labeled and unlabled strong augmentation."""
def prase_name(x):
if not unlabled2017_used:
return x + "-unlabeled"
else: # return coco2017 unlabeled data
return "coco_unlabeled2017"
def remove_no_box_data(_roidbs, filter_fn):
num = len(_roidbs)
_roidbs = filter_fn(_roidbs)
logger.info(
"Filtered {} images which contain no non-crowd groudtruth boxes. Total #images for training: {}"
.format(num - len(_roidbs), len(_roidbs)))
return _roidbs
pseudo_path = os.path.join(os.environ["PSEUDO_PATH"], "pseudo_data.npy")
pseudo_targets = dd.io.load(pseudo_path)
roidbs = list(
itertools.chain.from_iterable(
DatasetRegistry.get(x).training_roidbs() for x in cfg.DATA.TRAIN))
roidbs_u = list(
itertools.chain.from_iterable(
DatasetRegistry.get(prase_name(x)).training_roidbs()
for x in cfg.DATA.TRAIN))
roidbs = remove_no_box_data(
roidbs, lambda x: list(
filter(lambda img: len(img["boxes"][img["is_crowd"] == 0]) > 0, x))
)
roidbs_u = remove_no_box_data(
roidbs_u, lambda x: list(
filter(
lambda img: len(pseudo_targets[img["image_id"]]["boxes"]) > 0,
x)))
print_class_histogram(roidbs)
print_class_histogram(roidbs_u)
preprocess = TrainingDataPreprocessorSSlAug(
cfg, confidence=cfg.TRAIN.CONFIDENCE, pseudo_targets=pseudo_targets)
for jj, (rob, robu) in tqdm(enumerate(zip(roidbs, roidbs_u))):
data = preprocess((rob, robu))
# import pdb; pdb.set_trace()
nn = len(pseudo_targets[robu["image_id"]]["boxes"])
if data is None or len(data["gt_boxes_strong"]) == 0:
print("empty annotation, {} (original {})".format(jj, nn))
continue
ims = viz.draw_boxes(data["image"], data["gt_boxes"],
[str(a) for a in data["gt_labels"]])
ims_t = viz.draw_boxes(data["image_strong"], data["gt_boxes_strong"], [
str(a)
for a in data["gt_labels_strong"][:len(data["gt_boxes_strong"])]
])
ims = cv2.resize(ims, (ims_t.shape[1], ims_t.shape[0]))
vis = np.concatenate((ims, ims_t), axis=1)
if not os.path.exists(
os.path.dirname(
os.path.join(VISPATH, "result_{}.jpeg".format(jj)))):
os.makedirs(
os.path.dirname(
os.path.join(VISPATH, "result_{}.jpeg".format(jj))))
assert cv2.imwrite(os.path.join(VISPATH, "result_{}.jpeg".format(jj)),
vis)
if jj > maxvis:
break
for _ in tqdm(range(len(imagelist)),'Doing Predictions:'):
try:
_b,_s,_l=self.session.run([self.box,self.score,self.label])
obj = {'boxes': _b, 'labels': _l, 'scores': _s}
ret.append(obj)
except tf.errors.OutOfRangeError:
break
return ret
import tensorpack.utils.viz as viz
if __name__ == '__main__':
with open('../../data/coco.names') as fs:
names=fs.readlines()
print(names[0])
istraining=False
model_path='/home/zxk/AI/tensorpack/FRCNN/COCO-R50C4-MaskRCNN-Standard.npz'
service=FRCnnService(cfg,model_path)
imagelist=['../../data/demo_data/611.jpg']
result=service.predict_imagelist(imagelist)
im=cv2.imread(imagelist[0])
for r in result:
# print(r['boxes'].shape,r['labels'].shape,r['scores'].shape)
# print(r['scores'])
labels=['%s:%.2f'%(names[ll-1],round(ss,2)) for ll,ss in zip(r['labels'],r['scores'])]
print(r['labels'])
print(labels)
im=viz.draw_boxes(im,r['boxes'],labels)
viz.interactive_imshow(im)
def draw_final_outputs(img, results):
"""
Args:
results: [DetectionResult]
"""
# new_results = []
# for r in results:
# if r.score <=0.49:
# new_results.append(r)
# results = new_results
if len(results) == 0:
return img
# Display in largest to smallest order to reduce occlusion
boxes = np.asarray([r.box for r in results])
areas = np_area(boxes)
sorted_inds = np.argsort(-areas)
ret = img
tags = []
new_boxes = []
# rm_lst = class_nms(results, sorted_inds)
rm_lst = box_class_nms(results, sorted_inds)
print("rm_lst = ", rm_lst)
for result_id in sorted_inds:
if result_id in rm_lst:
continue
r = results[result_id]
# print("r = ", r)
if r.mask is not None:
level = str(r.class_id).split(" ")[0]
if "1" in level:
# color = (0, 255, 0)
# color = [0.000, 255.000, 0.000]
color_id = 23
# color_id = 9
elif "2" in level:
color_id = 22
# color_id = 9
# color = [0.000, 255.000, 255.000]
# color = (0, 255, 255)
elif "3" in level:
color_id = 9
# color = [0.000, 0.000, 255.000]
# color = (0, 0, 255)
else:
color = [0.000, 255.000, 0.000]
# color = (0, 255, 0)
print("error level!")
ret = draw_mask(ret, r.mask, color=None, color_id=color_id)
for result_id in sorted_inds:
if result_id in rm_lst:
continue
r = results[result_id]
new_boxes.append(r.box)
tags.append("{}, {:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id],
r.score))
# for r in results:
# tags.append(
# "{}, {:.2f}".format(cfg.DATA.CLASS_NAMES[r.class_id], r.score))
ret = viz.draw_boxes(ret, new_boxes, tags)
return ret
images_and_shape=list(map(common.processImage, imagelist))
images=np.array([im for im,_ in images_and_shape])
orgin_shape=[s for _,s in images_and_shape]
result=self.predict_416(images,batchSize,score_thresh,iou_thresh)
for r,orgin in zip(result,orgin_shape):
w,h=orgin
if len(r['boxes'])==0 :continue
bb=r['boxes']*np.array([w,h,w,h])/416
bb[:, [0, 2]] = np.maximum(bb[:, [0, 2]], 0)
bb[:, [1, 3]] = np.maximum(bb[:, [1, 3]], 0)
bb[:, [0, 2]] = np.minimum(bb[:, [0, 2]], w)
bb[:, [1, 3]] = np.minimum(bb[:, [1, 3]], h)
r['boxes']=bb
return result
import tensorpack.utils.viz as viz
if __name__ == '__main__':
model_path='/home/zxk/AI/tensorflow-yolov3/checkpoint/yolov3.ckpt'
service=YoLoService(model_path)
imagelist=['/home/zxk/PycharmProjects/deepAI1/daily/8/DeepLearning/myproject/yolo3/data/demo_data/611.jpg']
result=service.predict_imagelist(imagelist)
# for r in result:
# print(r['boxes'].shape,r['labels'].shape,r['scores'].shape)
# print(r['boxes'])
img=cv2.imread('/home/zxk/AI/coco/val2017/000000579893.jpg')
bbox=service.predict(img)
img=viz.draw_boxes(img,bbox['boxes'])
viz.interactive_imshow(img)