Mohammed Innat

@fchollet

I was referring something like as follows https://jax.readthedocs.io/en/latest/index.html

You can try `LabelMe`. https://github.com/wkentaro/labelme Or, another annotation tool developed by the author of the `ABCNet` model. It's a great ability to label the curve text very smoothly. You can...

With PyTorch is easy to implement the research idea you have. You can do it in Keras as well. It's a matter of preference.

@john-rocky Thanks for sharing the colab starter. Based on it, I've made an another demo starter on kaggle environment. https://www.kaggle.com/code/ipythonx/dichotomous-image-segmentation-inference

Thanks for your valuable comment. And thanks for the above link. However, it's actually hard to implement a precise function to get the weighted average of several polygons. However, here...

@ZFTurbo I don't have yet, but I can make it. However, how about following polygons from the above link that you gave! ``` shapes = [ {'type': 'Polygon', 'coordinates': [[(1095.76,...

Great. Now, it's midnight here. I will prepare something by tomorrow and share it with you. -)

@ZFTurbo I've picked 10 samples from the [Total-Text data](https://github.com/cs-chan/Total-Text-Dataset) data set. Inference on them using [MaskTextSpotterV3](https://github.com/MhLiao/MaskTextSpotterV3) model and saved 3 predicted outcomes of each (so total 30). This may not...

Only EfficietNet backbone. But can be integrated into another backbone easily. https://gist.github.com/innat/2d2fd2e93b47dd5349095182eb984ef6