How to apply mask_rcnn segmentation on a balloon video ? #101

gireeshkbogu · 2019-10-17T03:26:28Z

Hi, I am going through the google colab example tutorial.

I am trying to apply mask_rcnn segmentation on a random youtube balloon-video instead of an balloon-image to detect balloon only (one class).

How can I assign .yaml and .pkl files that were generated using images earlier in the tutorial to a random video? thanks

I tried the foolowing but it didn't work. I think I am having trouble assign the trained config and model files.

!cd detectron2_repo && python demo/demo.py --config-file configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml --video-input ../video-clip_b.mp4 --confidence-threshold 0.6 --output ../video-clip_b_testing1.mkv \
  --opts MODEL.WEIGHTS detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl

The text was updated successfully, but these errors were encountered:

ppwwyyxx · 2019-10-17T03:52:17Z

You can find the following config in the tutorial, which is used to train the balloon model:

cfg.DATASETS.TRAIN = ("balloon/train",)
cfg.DATASETS.TEST = ()   # no metrics implemented for this dataset
cfg.DATALOADER.NUM_WORKERS = 2
cfg.MODEL.WEIGHTS = "detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl"  # initialize from model zoo
cfg.SOLVER.IMS_PER_BATCH = 2
cfg.SOLVER.BASE_LR = 0.00025
cfg.SOLVER.MAX_ITER = 300    # 300 iterations seems good enough, but you can certainly train longer
cfg.MODEL.ROI_HEADS.BATCH_SIZE_PER_IMAGE = 128   # faster, and good enough for this toy dataset
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 1  # only has one class (ballon)

the config file you're running demo.py with does not have these changes, so it won't work.

The model you trained will be in os.path.join(cfg.OUTPUT_DIR, "model_final.pth"), as you can find in the tutorial as well. You need to set MODEL.WEIGHTS to it for inference.

gireeshkbogu · 2019-10-17T03:58:40Z

I tried to do this but didn't work. Is this because ballon_dataset was directed to /dev/null in the tutorial and somehow this is messing with the paths?

cfg.MODEL.WEIGHTS = os.path.join(cfg.OUTPUT_DIR, "model_final.pth")

!cd detectron2_repo && python demo/demo.py --config-file configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml --video-input ../video-clip_b.mp4 --confidence-threshold 0.6 --output ../video-clip_b_testing1.mkv \
  --opts MODEL.WEIGHTS cfg.MODEL.WEIGHTS

Error

[10/17 04:02:19 detectron2]: Arguments: Namespace(confidence_threshold=0.6, config_file='configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml', input=None, opts=['MODEL.WEIGHTS', 'cfg.MODEL.WEIGHTS'], output='../video-clip_b_testing1.mkv', video_input='../video-clip_b.mp4', webcam=False)
WARNING [10/17 04:02:19 d2.config.compat]: Config 'configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
Traceback (most recent call last):
  File "demo/demo.py", line 73, in <module>
    demo = VisualizationDemo(cfg)
  File "/content/detectron2_repo/demo/predictor.py", line 33, in __init__
    self.predictor = DefaultPredictor(cfg)
  File "/content/detectron2_repo/detectron2/engine/defaults.py", line 149, in __init__
    checkpointer.load(cfg.MODEL.WEIGHTS)
  File "/usr/local/lib/python3.6/dist-packages/fvcore/common/checkpoint.py", line 99, in load
    assert os.path.isfile(path), "Checkpoint {} not found!".format(path)
AssertionError: Checkpoint cfg.MODEL.WEIGHTS not found!

ppwwyyxx · 2019-10-17T04:29:52Z

cfg.MODEL.WEIGHTS is a python statement. The second line is a shell command. No python variables will be available in the shell command.

ppwwyyxx · 2019-10-17T04:33:09Z

Similarly, all the changes you made to cfg affect the python variable. It does not affect the shell command you run, which reads config from the given config file.

gireeshkbogu · 2019-10-17T04:47:44Z

Ok now it is running but the output is the same as input with no segmentation at all.

!cd detectron2_repo && python demo/demo.py --config-file configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml --video-input ../video-clip_b.mp4 --confidence-threshold 0.6 --output ../video-clip_b_testing2.mkv \
  --opts MODEL.WEIGHTS ../output/model_final.pth

[10/17 04:49:22 detectron2]: Arguments: Namespace(confidence_threshold=0.7, config_file='configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml', input=None, opts=['MODEL.WEIGHTS', '../output/model_final.pth'], output='../video-clip_b_testing4.mkv', video_input='../video-clip_b.mp4', webcam=False)
WARNING [10/17 04:49:22 d2.config.compat]: Config 'configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
'roi_heads.box_predictor.cls_score.weight' has shape (2, 1024) in the checkpoint but (81, 1024) in the model! Skipped.
'roi_heads.box_predictor.cls_score.bias' has shape (2,) in the checkpoint but (81,) in the model! Skipped.
'roi_heads.box_predictor.bbox_pred.weight' has shape (4, 1024) in the checkpoint but (320, 1024) in the model! Skipped.
'roi_heads.box_predictor.bbox_pred.bias' has shape (4,) in the checkpoint but (320,) in the model! Skipped.
'roi_heads.mask_head.predictor.weight' has shape (1, 256, 1, 1) in the checkpoint but (80, 256, 1, 1) in the model! Skipped.
'roi_heads.mask_head.predictor.bias' has shape (1,) in the checkpoint but (80,) in the model! Skipped.
100% 181/181 [01:28<00:00,  2.04it/s]

ppwwyyxx · 2019-10-17T05:35:55Z

all the changes you made to cfg affect the python variable. It does not affect the shell command you run, which reads config from the given config file.

Which means you need to change the config file given to demo.py.

gireeshkbogu · 2019-10-18T16:55:21Z

to something like this? I ended up with the following error.
Do I need to for the detectron2 rep and change the yaml file instead?

_BASE_: "../Base-RCNN-FPN.yaml"
MODEL:
  WEIGHTS: "../output/model_final.pth"
  MASK_ON: True
  RESNETS:
    DEPTH: 50
SOLVER:
  STEPS: (210000, 250000)
  MAX_ITER: 270000

Error

Failed to load OpenCL runtime
[10/18 17:12:25 detectron2]: Arguments: Namespace(confidence_threshold=0.7, config_file='/content/gdrive/My Drive/fastai_files/datasets/test_samples/balloon.yaml', input=None, opts=['MODEL.WEIGHTS', '../output/model_final.pth'], output='../video-clip_bb_testing4.mkv', video_input='../video-clip_bb.mp4', webcam=False)
Traceback (most recent call last):
  File "demo/demo.py", line 71, in <module>
    cfg = setup_cfg(args)
  File "demo/demo.py", line 23, in setup_cfg
    cfg.merge_from_file(args.config_file)
  File "/content/detectron2_repo/detectron2/config/config.py", line 23, in merge_from_file
    loaded_cfg = _CfgNode.load_yaml_with_base(cfg_filename, allow_unsafe=allow_unsafe)
  File "/usr/local/lib/python3.6/dist-packages/fvcore/common/config.py", line 75, in load_yaml_with_base
    if BASE_KEY in cfg:
TypeError: argument of type 'NoneType' is not iterable

ppwwyyxx · 2019-10-18T18:33:53Z

You need to create a yaml file which has the same content as configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml, except for the extra changes mentioned above in #101 (comment).

gireeshkbogu · 2019-10-18T19:44:39Z

I think that is what I did. I am writing all the steps just to be clear.

First I ran this to create config on balloon train samples

from detectron2.engine import DefaultTrainer
from detectron2.config import get_cfg

cfg = get_cfg()
cfg.merge_from_file("./detectron2_repo/configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml")
cfg.DATASETS.TRAIN = ("balloon/train",)
cfg.DATASETS.TEST = ()   # no metrics implemented for this dataset
cfg.DATALOADER.NUM_WORKERS = 2
cfg.MODEL.WEIGHTS = "detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl"  # initialize from model zoo
cfg.SOLVER.IMS_PER_BATCH = 2
cfg.SOLVER.BASE_LR = 0.00025
cfg.SOLVER.MAX_ITER = 300    # 300 iterations seems good enough, but you can certainly train longer
cfg.MODEL.ROI_HEADS.BATCH_SIZE_PER_IMAGE = 128   # faster, and good enough for this toy dataset
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 1  # only has one class (ballon)

os.makedirs(cfg.OUTPUT_DIR, exist_ok=True)
trainer = DefaultTrainer(cfg) 
trainer.resume_or_load(resume=False)
trainer.train()

configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml looks like the following ..

_BASE_: "../Base-RCNN-FPN.yaml"
MODEL:
  WEIGHTS: "detectron2://ImageNetPretrained/MSRA/R-50.pkl"
  MASK_ON: True
  RESNETS:
    DEPTH: 50
SOLVER:
  STEPS: (210000, 250000)
  MAX_ITER: 270000

So I manually created a yaml file like this in a different folder (/content/gdrive/My Drive/fastai_files/datasets/test_samples/balloon.yaml).

_BASE_: "../Base-RCNN-FPN.yaml"
MODEL:
  WEIGHTS: "../output/model_final.pth"
  MASK_ON: True
  RESNETS:
    DEPTH: 50
SOLVER:
  STEPS: (210000, 250000)
  MAX_ITER: 270000

and then finally ran like the command using my own yaml file and model weights

!cd detectron2_repo && python demo/demo.py --config-file '/content/gdrive/My Drive/fastai_files/datasets/test_samples/balloon.yaml' --video-input ../video-clip_bb.mp4 --confidence-threshold 0.7 --output ../video-clip_bb_testing_1.mkv \
  --opts MODEL.WEIGHTS ../output/model_final.pth

ppwwyyxx · 2019-10-18T21:34:28Z

all the changes you made to cfg affect the python variable. It does not affect the shell command you run, which reads config from the given config file.

What the command run is demo.py. Running demo.py has nothing to with the changes

cfg = get_cfg()
cfg.merge_from_file("./detectron2_repo/configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml")
cfg.DATASETS.TRAIN = ("balloon/train",)
cfg.DATASETS.TEST = ()   # no metrics implemented for this dataset
cfg.DATALOADER.NUM_WORKERS = 2
cfg.MODEL.WEIGHTS = "detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl"  # initialize from model zoo
cfg.SOLVER.IMS_PER_BATCH = 2
cfg.SOLVER.BASE_LR = 0.00025
cfg.SOLVER.MAX_ITER = 300    # 300 iterations seems good enough, but you can certainly train longer
cfg.MODEL.ROI_HEADS.BATCH_SIZE_PER_IMAGE = 128   # faster, and good enough for this toy dataset
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 1  # only has one class (ballon)

that you made during training, because these changes are not in demo.py, and not in the config file you gave to demo.py either.

You need to add these changes either to demo.py, or to the config file you provided to demo.py.

gireeshkbogu · 2019-10-18T22:51:22Z

This time I forked Detectron2 repo and re-ran all the code by adding config changes to the but still have issues.

!cd detectron2_repo && python demo/demo_balloon.py --config-file 'configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml' --video-input ../video-clip_bb.mp4 --confidence-threshold 0.7 --output ../video-clip_bb_testing_1.mkv \
  --opts MODEL.WEIGHTS ../output/model_final.pth

Failed to load OpenCL runtime
[10/18 22:35:43 detectron2]: Arguments: Namespace(confidence_threshold=0.7, config_file='configs/quick_schedules/e2e_mask_rcnn_R_50_FPN_inference_acc_test.yaml', input=None, opts=['MODEL.WEIGHTS', '../output/model_final.pth'], output='../video-clip_bb_testing_1.mkv', video_input='../video-clip_bb.mp4', webcam=False)
Traceback (most recent call last):
  File "demo/demo_balloon.py", line 79, in <module>
    cfg = setup_cfg(args)
  File "demo/demo_balloon.py", line 23, in setup_cfg
    cfg.merge_from_file("./detectron2_repo/configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml")
  File "/content/detectron2_repo/detectron2/config/config.py", line 23, in merge_from_file
    loaded_cfg = _CfgNode.load_yaml_with_base(cfg_filename, allow_unsafe=allow_unsafe)
  File "/usr/local/lib/python3.6/dist-packages/fvcore/common/config.py", line 47, in load_yaml_with_base
    with open(filename, "r") as f:
FileNotFoundError: [Errno 2] No such file or directory: './detectron2_repo/configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml'

ppwwyyxx · 2019-10-18T22:57:00Z

FileNotFoundError: [Errno 2] No such file or directory: './detectron2_repo/configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml'

As the error says, your code is trying to read a file that does not exist. Since your command has "cd detectron2_repo", you should remove "detectron2_repo" from the path.

gireeshkbogu · 2019-10-18T23:14:12Z

Thank you now having the following issue

!cd detectron2_repo && python demo/demo_balloon.py --video-input ../video-clip_bb.mp4 --confidence-threshold 0.7 --output ../video-clip_bb_testing_1.mkv \
  --opts MODEL.WEIGHTS ../output/model_final.pth

Failed to load OpenCL runtime
[10/18 23:13:30 detectron2]: Arguments: Namespace(confidence_threshold=0.7, config_file='configs/quick_schedules/e2e_mask_rcnn_R_50_FPN_inference_acc_test.yaml', input=None, opts=['MODEL.WEIGHTS', '../output/model_final.pth'], output='../video-clip_bb_testing_1.mkv', video_input='../video-clip_bb.mp4', webcam=False)
WARNING [10/18 23:13:30 d2.config.compat]: Config 'configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
Traceback (most recent call last):
  File "demo/demo_balloon.py", line 81, in <module>
    demo = VisualizationDemo(cfg)
  File "/content/detectron2_repo/demo/predictor.py", line 24, in __init__
    self.metadata = MetadataCatalog.get(cfg.DATASETS.TEST[0])
IndexError: tuple index out of range

ppwwyyxx · 2019-10-18T23:20:49Z

Set cfg.DATASETS.TEST = ("balloon/val",)
// UPDATE: or balloon_val for latest tutorial

gireeshkbogu · 2019-10-18T23:40:09Z

Thank you. Seems to be running but got stuck here

Failed to load OpenCL runtime
[10/18 23:37:22 detectron2]: Arguments: Namespace(confidence_threshold=0.7, config_file='configs/quick_schedules/e2e_mask_rcnn_R_50_FPN_inference_acc_test.yaml', input=None, opts=['MODEL.WEIGHTS', '../output/model_final.pth'], output='../video-clip_bb_testing_1.mkv', video_input='../video-clip_bb.mp4', webcam=False)
WARNING [10/18 23:37:22 d2.config.compat]: Config 'configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
'roi_heads.box_predictor.cls_score.weight' has shape (81, 1024) in the checkpoint but (2, 1024) in the model! Skipped.
'roi_heads.box_predictor.cls_score.bias' has shape (81,) in the checkpoint but (2,) in the model! Skipped.
'roi_heads.box_predictor.bbox_pred.weight' has shape (320, 1024) in the checkpoint but (4, 1024) in the model! Skipped.
'roi_heads.box_predictor.bbox_pred.bias' has shape (320,) in the checkpoint but (4,) in the model! Skipped.
'roi_heads.mask_head.predictor.weight' has shape (80, 256, 1, 1) in the checkpoint but (1, 256, 1, 1) in the model! Skipped.
'roi_heads.mask_head.predictor.bias' has shape (80,) in the checkpoint but (1,) in the model! Skipped.
  5% 9/181 [00:04<01:26,  1.99it/s]Traceback (most recent call last):
  File "demo/demo_balloon.py", line 145, in <module>
    for vis_frame in tqdm.tqdm(demo.run_on_video(video), total=num_frames):
  File "/usr/local/lib/python3.6/dist-packages/tqdm/std.py", line 1081, in __iter__
    for obj in iterable:
  File "/content/detectron2_repo/demo/predictor.py", line 127, in run_on_video
    yield process_predictions(frame, self.predictor(frame))
  File "/content/detectron2_repo/demo/predictor.py", line 96, in process_predictions
    vis_frame = video_visualizer.draw_instance_predictions(frame, predictions)
  File "/content/detectron2_repo/detectron2/utils/video_visualizer.py", line 90, in draw_instance_predictions
    labels = _create_text_labels(classes, scores, self.metadata.thing_classes)
  File "/content/detectron2_repo/detectron2/data/catalog.py", line 116, in __getattr__
    key, self.name, str(self.__dict__.keys())
AttributeError: Attribute 'thing_classes' does not exist in the metadata of 'balloon/val'. Available keys are dict_keys(['name']).
  5% 9/181 [00:05<01:37,  1.77it/s]

ppwwyyxx · 2019-10-18T23:48:49Z

This error is addressed by 22e04d1

gireeshkbogu · 2019-10-18T23:51:33Z

If I have to add this new change do I need to repeat the forking/install and rerun things?

ppwwyyxx · 2019-10-18T23:56:51Z

If you install detectron2 in the same way as the tutorial, then git pull in the repo is sufficient.

gireeshkbogu · 2019-10-19T00:20:23Z

All ran well but the output is the same as the input. There was no object detection in the video.

Do you think this could be a threshold score issue? When I change it from 0.7 to 0.6, there were boxes popping everywhere but not where the balloons are. This is strange.

Failed to load OpenCL runtime
[10/19 00:15:48 detectron2]: Arguments: Namespace(confidence_threshold=0.7, config_file='configs/quick_schedules/e2e_mask_rcnn_R_50_FPN_inference_acc_test.yaml', input=None, opts=['MODEL.WEIGHTS', '../output/model_final.pth'], output='../video-clip_bb_testing_1.mkv', video_input='../video-clip_bb.mp4', webcam=False)
WARNING [10/19 00:15:48 d2.config.compat]: Config 'configs/COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x.yaml' has no VERSION. Assuming it to be compatible with latest v2.
'roi_heads.box_predictor.cls_score.weight' has shape (81, 1024) in the checkpoint but (2, 1024) in the model! Skipped.
'roi_heads.box_predictor.cls_score.bias' has shape (81,) in the checkpoint but (2,) in the model! Skipped.
'roi_heads.box_predictor.bbox_pred.weight' has shape (320, 1024) in the checkpoint but (4, 1024) in the model! Skipped.
'roi_heads.box_predictor.bbox_pred.bias' has shape (320,) in the checkpoint but (4,) in the model! Skipped.
'roi_heads.mask_head.predictor.weight' has shape (80, 256, 1, 1) in the checkpoint but (1, 256, 1, 1) in the model! Skipped.
'roi_heads.mask_head.predictor.bias' has shape (80,) in the checkpoint but (1,) in the model! Skipped.
100% 181/181 [01:32<00:00,  1.97it/s]

ppwwyyxx · 2019-10-19T02:05:36Z

As I said in #101 (comment)

The model you trained will be in os.path.join(cfg.OUTPUT_DIR, "model_final.pth"), as you can find in the tutorial as well. You need to set MODEL.WEIGHTS to it for inference.

You load a wrong model.

gireeshkbogu · 2019-10-19T02:37:38Z

In the tutorial .pkl file was set to weights. Setting .pth is the same as .pkl?
Also when I build the model with balloon data, where does it save the .pkl file?

ppwwyyxx · 2019-10-19T02:40:37Z

In the tutorial .pkl file was set to weights. Setting .pth is the same as .pkl?

Both file extensions are supported.

Also when I build the model with balloon data, where does it save the .pkl file?

I do not know what you mean by "when I build the model", and when saying "the .pkl file" it is unclear to me what pkl file you're referring to.

gireeshkbogu · 2019-10-19T02:42:52Z

Here is the line from the tutorial where *.pkl file set to model weights in the demo.py

#cfg.MODEL.WEIGHTS = "detectron2://COCO-InstanceSegmentation/mask_rcnn_R_50_FPN_3x/137849600/model_final_f10217.pkl"

If I understand correctly, I just replace this .pkl with my own .pth file in demo.py?

ppwwyyxx · 2019-10-19T02:47:54Z

Yes

gireeshkbogu · 2019-10-19T03:04:14Z

Yey!!! it is working. Thank you for being so patient with me.
I really really appreciate your help even at this time. I hope you have a nice weekend :)

Haow235 · 2020-01-02T17:57:24Z

Set cfg.DATASETS.TEST = ("balloon/val",)

@ppwwyyxx

I trained just like the balloon tutorial, got the trained result, and I have the same error when predict in my own new dataset, which doesn't have label, only image, I think it's because I don't set the config correctly, would you please tell me how to set the cfg correctly when predict in new dataset.
The following is my code:
`from google.colab.patches import cv2_imshow

cfg.MODEL.WEIGHTS = os.path.join(cfg.OUTPUT_DIR, "model_final.pth")
cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.7 # set the testing threshold for this model
cfg.MODEL.ROI_HEADS.NUM_CLASSES = 6 # only has one class (ballon)
predictor = DefaultPredictor(cfg)
outputs = predictor(im) IndexError Traceback (most recent call last)

in ()
6 cfg.MODEL.ROI_HEADS.NUM_CLASSES = 6 # only has one class (ballon)
7 # cfg.DATASETS.TEST = ("balloon_val", )
----> 8 predictor = DefaultPredictor(cfg)
9 outputs = predictor(im)

/content/detectron2_repo/detectron2/engine/defaults.py in init(self, cfg)
157 self.model = build_model(self.cfg)
158 self.model.eval()
--> 159 self.metadata = MetadataCatalog.get(cfg.DATASETS.TEST[0])
160
161 checkpointer = DetectionCheckpointer(self.model)

IndexError: tuple index out of range`

anyhelp would be appreciate, thank you so much!

ppwwyyxx · 2020-01-02T18:15:03Z

#101 (comment)

Haow235 · 2020-01-02T22:34:06Z

thank you, I figured it out. I need to register the test data before using it, and then do as you say. thanks a lot~

ppwwyyxx added the usage label Oct 17, 2019

gireeshkbogu closed this as completed Oct 19, 2019

EchizenG mentioned this issue Dec 26, 2019

No segmentation when inferencing #585

Closed

github-actions bot locked as resolved and limited conversation to collaborators Jan 6, 2021

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

How to apply mask_rcnn segmentation on a balloon video ? #101

How to apply mask_rcnn segmentation on a balloon video ? #101

gireeshkbogu commented Oct 17, 2019 •

edited

ppwwyyxx commented Oct 17, 2019

gireeshkbogu commented Oct 17, 2019 •

edited

ppwwyyxx commented Oct 17, 2019

ppwwyyxx commented Oct 17, 2019

gireeshkbogu commented Oct 17, 2019 •

edited

ppwwyyxx commented Oct 17, 2019

gireeshkbogu commented Oct 18, 2019 •

edited

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019 •

edited

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019 •

edited

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019 •

edited

ppwwyyxx commented Oct 18, 2019 •

edited

gireeshkbogu commented Oct 18, 2019

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 19, 2019 •

edited

ppwwyyxx commented Oct 19, 2019

gireeshkbogu commented Oct 19, 2019

ppwwyyxx commented Oct 19, 2019

gireeshkbogu commented Oct 19, 2019 •

edited

ppwwyyxx commented Oct 19, 2019

gireeshkbogu commented Oct 19, 2019

Haow235 commented Jan 2, 2020 •

edited

ppwwyyxx commented Jan 2, 2020

Haow235 commented Jan 2, 2020

How to apply mask_rcnn segmentation on a balloon video ? #101

How to apply mask_rcnn segmentation on a balloon video ? #101

Comments

gireeshkbogu commented Oct 17, 2019 • edited

ppwwyyxx commented Oct 17, 2019

gireeshkbogu commented Oct 17, 2019 • edited

ppwwyyxx commented Oct 17, 2019

ppwwyyxx commented Oct 17, 2019

gireeshkbogu commented Oct 17, 2019 • edited

ppwwyyxx commented Oct 17, 2019

gireeshkbogu commented Oct 18, 2019 • edited

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019 • edited

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019 • edited

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019 • edited

ppwwyyxx commented Oct 18, 2019 • edited

gireeshkbogu commented Oct 18, 2019

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 18, 2019

ppwwyyxx commented Oct 18, 2019

gireeshkbogu commented Oct 19, 2019 • edited

ppwwyyxx commented Oct 19, 2019

gireeshkbogu commented Oct 19, 2019

ppwwyyxx commented Oct 19, 2019

gireeshkbogu commented Oct 19, 2019 • edited

ppwwyyxx commented Oct 19, 2019

gireeshkbogu commented Oct 19, 2019

Haow235 commented Jan 2, 2020 • edited

ppwwyyxx commented Jan 2, 2020

Haow235 commented Jan 2, 2020

gireeshkbogu commented Oct 17, 2019 •

edited

gireeshkbogu commented Oct 17, 2019 •

edited

gireeshkbogu commented Oct 17, 2019 •

edited

gireeshkbogu commented Oct 18, 2019 •

edited

gireeshkbogu commented Oct 18, 2019 •

edited

gireeshkbogu commented Oct 18, 2019 •

edited

gireeshkbogu commented Oct 18, 2019 •

edited

ppwwyyxx commented Oct 18, 2019 •

edited

gireeshkbogu commented Oct 19, 2019 •

edited

gireeshkbogu commented Oct 19, 2019 •

edited

Haow235 commented Jan 2, 2020 •

edited