waspinator

it would also be good to update the documentation for using v2.1. For example the [installation](https://github.com/burzum/cakephp-file-storage/blob/2.1.0/docs/Documentation/Installation.md) page references `use Burzum\FileStorage\Lib\FileStorageUtils;` which I don't think exists anymore.

So if I add an email to the queue with these options ``` $to = '[email protected]'; $view_vars = [ 'variable_name' => $variable_value ]; $send_at = Time::now(); $options = [ 'subject'...

There are a few annotators out there, but two I know are geared to COCO are - https://github.com/uoguelph-ri/coco-annotator - https://github.com/waspinator/js-segment-annotator

The easiest way to get started is by trying out the notebook in the example. https://jupyter.readthedocs.io/en/latest/install.html

Sorry, not directly with this tool. If you can create binary numpy masks from the json data, then you can convert to COCO that way.

https://github.com/waspinator/js-segment-annotator there is a python tool to help convert the output to coco

you can decode RLE into binary masks using the [decode](https://github.com/cocodataset/cocoapi/blob/master/PythonAPI/pycocotools/mask.py#L32) function in the original cocoapi tools, and then use [binary_mask_to_polygon](https://github.com/waspinator/pycococreator/blob/master/pycococreatortools/pycococreatortools.py#L35) to get a polygon. But the current coco format doesn't...

currently the coco format does not support holes in object annotations

Everything needed to train should be there. I'll be publishing an article on how to use Matterport's Mask RCNN implementation on the shapes dataset next week, which should hopefully answer...

can you share the original image and the binary mask of the person, so I can try to reproduce the issue?