DeepForest
DeepForest copied to clipboard
weecology
Python Package for Airborne RGB machine learning
Issue #569 fixed
Currently the output method is a simple plot and `boxes.head()` will give `x_min` `x_max` `y_min` `y_max` bounding boxes around detections. It would be great if, in the case the input...
Fixes Issue #516: The following modifications have been made: - Updated the `boxes_to_shapefile` function in `utilities.py` to use the 'rgb' argument instead of 'image_path'. - Modified the test functions in...
https://openaccess.thecvf.com/content_CVPRW_2020/papers/w11/Li_Density_Map_Guided_Object_Detection_in_Aerial_Images_CVPRW_2020_paper.pdf We often have multi-scale challenges, especially as predict_tile just cuts images into uniform pieces. This an interesting idea for bird detection. with trees the images are often uniformly trees.
When we are dealing with multi-class models, we wait pretty late until a non-matching label error is thrown. ``` Exception has occurred: KeyError Caught KeyError in DataLoader worker process 0....
See #394. It's a little confusing because all of the layers that we trained are being fine tuned, but (I think) not all of the layers in the underlying Retinanet.
I've noticed that ``` from deepforest import main ``` is quite slow, much slower than any typical package. What is causing this?
predict.predict_file needs a resizing function.
It is more a question, I would like to understand why the training does not use the background as a class equal to 0. The documentation of RetinaNet stipulates: num_classes...
Useful to show people to use create_trainer to pip in kwargs.