DeepTreeAttention
DeepTreeAttention copied to clipboard
weecology
Implementation of Hang et al. 2020 "Hyperspectral Image Classification with Attention Aided CNNs" for tree species prediction
rasterio hates multiprocessing, we have enough GPU to just load it up
At test time we can use segment anything but on deepforest box.
https://web.eecs.umich.edu/~stellayu/publication/doc/2023hierSPL.pdf talk to Zhongqi Miao
https://arxiv.org/abs/2305.13210 I believe there was a recent tree paper here as well.
https://drive.google.com/file/d/1Ei1z2tbpmjjB197ZNVCBxOB4S9fUr-xH/view
BRDF
Create a branch that computes updated reflectance based on azimuth and sun angle before cropping.
The improvements in padding may help RGB texture feature learning.
Get a slice of a small number of crowns to see a rgb band combo
We need to get the rgb visualizations back into the confusion matrix, check the main branch for code under visualize.py