Janne Mäyrä

Hi, The workflow is pretty much the following: 1. Check the raw hyperspectral data and see if it needs some processing, mostly covered in [Data exploration and preprocessing.ipynb](https://github.com/jaeeolma/tree-detection-evo/blob/master/notebooks/Data%20exploration%20and%20preprocessing.ipynb). In our...

`train_stats.npy` is a file that contains both mean and std for training data, used for normalizing the data. You have to get those for your own data, as unfortunately hyperspectral...

Finally got time to work on this. Only breaking change that I found is that each time we use `detection_model.get_prediction()` or `detection_model.convert_original_predictions` with mask output, `full_shape` must be included as...

@Preburk good point, probably should try to implement NMS for polygons. Using shapely should work cause IoU is just ```python def poly_IoU(poly_1:Polygon, poly_2: Polygon) -> float: area_intersection = poly_1.intersection(poly_2).area area_union...

Continues in #1039

This is slightly confusing, because I get different results with `results = model('path/to/image.jpg')` and ```python im_bgr = cv2.imread('path/to/image.jpg') im_rgb = cv2.cvtColor(im_bgr, cv2.COLOR_BGR2RGB) results = model(im_rgb) ``` However, predictions between using...

This actually doesn't clear up things, because as seen in the example codes, the following produce identical results: ``` image_fn = 'path/to/image.jpg' im_bgr = cv2.imread(image_fn) preds_with fname = model(image_fn) #...