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

Mask RCNN and RLE #22

Open
enemni opened this issue Apr 25, 2019 · 3 comments
Open

Mask RCNN and RLE #22

enemni opened this issue Apr 25, 2019 · 3 comments

Comments

@enemni
Copy link

enemni commented Apr 25, 2019

Hello everyone,

I am running the matterport Mask_RCNN on my own dataset created using waspinator/pycococreator

Everything works fine when I am using polygons, but as soon as I am using RLE because I have crowded images I got an error when trying to display a few images from the training dataset,

I got problems with images with holes using polygons and I assumed that is why the pycococreator uses RLE when the image is crowded. This part works perfectly, but then since MaskRCCN requires polygons I am stuck in this loop.

Is there a way to convert RLE to Polygons? I am not even sure this would solve the problem

@waspinator
Copy link
Owner

waspinator commented May 29, 2019

you can decode RLE into binary masks using the decode function in the original cocoapi tools, and then use binary_mask_to_polygon to get a polygon.

But the current coco format doesn't encode holes in objects properly, so this may not help solve your issue.

@gethubwy
Copy link

gethubwy commented Jul 5, 2020

Hello everyone,

I am running the matterport Mask_RCNN on my own dataset created using waspinator/pycococreator

Everything works fine when I am using polygons, but as soon as I am using RLE because I have crowded images I got an error when trying to display a few images from the training dataset,

I got problems with images with holes using polygons and I assumed that is why the pycococreator uses RLE when the image is crowded. This part works perfectly, but then since MaskRCCN requires polygons I am stuck in this loop.

Is there a way to convert RLE to Polygons? I am not even sure this would solve the problem

Hello,Excuse me, I want to know how to train a new mode, I use the software named "labelme" to label my data, I want to train a new model starting from pre-trained COCO weights, but I don't know how to write the annotation file in validate set, are there tools to help me? Thank you.

@enemni
Copy link
Author

enemni commented Jul 14, 2020

Hello @gethubwy, I am not sure I get your question, can you please share the code with us. labelme should solve your problem and label your entire dataset

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

No branches or pull requests

3 participants