Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Update new bbs coords using imaug functions #311

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

alice17
Copy link

@alice17 alice17 commented Jun 20, 2018

No description provided.

@bkanaki
Copy link

bkanaki commented Sep 1, 2018

Hey, I tried to use your patch. Did you ever encounter this error:

File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\site-packages\tensorflow\python\keras\utils\data_utils.py", line 607, in get inputs = self.queue.get(block=True).get() File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\multiprocessing\pool.py", line 644, in get raise self._value File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\multiprocessing\pool.py", line 119, in worker result = (True, func(*args, **kwds)) File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\site-packages\tensorflow\python\keras\utils\data_utils.py", line 432, in get_index return _SHARED_SEQUENCES[uid][i] File "C:\Users\bhargav\Documents\code_base\ped_det\model\preprocessor.py", line 149, in __getitem__ img, all_objs = self.aug_image(train_instance, jitter=self.jitter) File "C:\Users\bhargav\Documents\code_base\ped_det\model\preprocessor.py", line 243, in aug_image bbs_aug = seq_det.augment_bounding_boxes([bbs])[0] File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\site-packages\imgaug\augmenters\meta.py", line 686, in augment_bounding_boxes y2=y2 File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\site-packages\imgaug\imgaug.py", line 1524, in copy label=self.label if label is None else label File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\site-packages\imgaug\imgaug.py", line 1259, in __init__ do_assert(x2 > x1) File "C:\Users\bhargav\AppData\Local\Continuum\miniconda3\envs\tf110gpu\lib\site-packages\imgaug\imgaug.py", line 678, in do_assert raise AssertionError(str(message)) AssertionError: Assertion failed.

@alice17
Copy link
Author

alice17 commented Sep 4, 2018

I think something is wrong with the coordinates of your bounding boxes: x2 must be greater than x1.

@bkanaki
Copy link

bkanaki commented Sep 6, 2018

I understand that from the error message. However, I am using coco 2017 dataset and I don't think there is any error in the bounding boxes for coco dataset as far as I know. In fact, I subtracted 1 and added 1 from min and max values respectively. I still get the same error, after few epochs!

@bkanaki
Copy link

bkanaki commented Sep 10, 2018

I am not sure what creates this error. Here is the list of bounding boxes when this error is raised:

BoundingBoxesOnImage([
    BoundingBox(x1=509.0000, y1=63.0000, x2=628.0000, y2=252.0000, label=None), 
    BoundingBox(x1=382.0000, y1=119.0000, x2=564.0000, y2=254.0000, label=None), 
    BoundingBox(x1=467.0000, y1=44.0000, x2=498.0000, y2=123.0000, label=None), 
    BoundingBox(x1=579.0000, y1=22.0000, x2=640.0000, y2=129.0000, label=None), 
    BoundingBox(x1=188.0000, y1=52.0000, x2=351.0000, y2=244.0000, label=None), 
    BoundingBox(x1=352.0000, y1=18.0000, x2=406.0000, y2=150.0000, label=None), 
    BoundingBox(x1=288.0000, y1=27.0000, x2=344.0000, y2=149.0000, label=None), 
    BoundingBox(x1=228.0000, y1=18.0000, x2=274.0000, y2=142.0000, label=None), 
    BoundingBox(x1=180.0000, y1=23.0000, x2=227.0000, y2=67.0000, label=None), 
    BoundingBox(x1=126.0000, y1=20.0000, x2=185.0000, y2=149.0000, label=None), 
    BoundingBox(x1=52.0000, y1=18.0000, x2=128.0000, y2=150.0000, label=None), 
    BoundingBox(x1=3.0000, y1=20.0000, x2=61.0000, y2=138.0000, label=None), 
    BoundingBox(x1=392.0000, y1=4.0000, x2=437.0000, y2=43.0000, label=None), 
    BoundingBox(x1=28.0000, y1=5.0000, x2=397.0000, y2=47.0000, label=None
)], shape=(278, 640, 3))

@rodrigo2019
Copy link

@alice17 and @bkanaki I did a similar implementation and it looks working fine to me.
check here

@bkanaki
Copy link

bkanaki commented Sep 11, 2018

Would the KeyPoints v.s. BoundingBox create an error? Theoretically, it shouldn't if the implementation for both are very similar in terms of sanity check.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants