IR-imaging-2024

Repository for the semester project carried out by Octavio Profeta in the spring of 2024, under the supervision of Pr. Francesco Mondada and Cyril Monette

training images for the ilastik model

hive2_rpi1_240423-034701Z hive2_rpi2_240425-164102Z hive2_rpi3_240423-020502Z hive2_rpi4_240424-014501Z

Marche à suivre

All the necessary dataset is in the live_bees folder

If you don't have all annotation (live_bees/rpi/masks/.csv):

0. Run the napari_annotation/pipeline.ipynb notebook until the indicated mark to start annotating. You need to do that for each missing RPi's masks
   • Once done with one RPi, run the remaining cells to save the masks
   • Rerun the notebook for each missing RPi's masks

If you have all the masks:

1. Run napari_annotation/csv_to_mask.ipynb notebook for each RPi
2. Run napari_annotation/csv_to_contour.ipynb notebook for each RPi

You can now start to run the mask finding pipelines:

3. Run the aa_thresholding/thresholding.ipynb notebook for each RPi. The resulting masks are in a_processed_images/thresholding folder
4. Run the aa_region_growing/region_growing.ipynb notebook for each RPi. The resulting masks are in a_processed_images/region_growing folder
5. Run the aa_ilastik/ilastik.ipynb notebook for each RPi. The resulting masks are in a_processed_images/ilastik folder
6. Run the aa_optical_flow/optical_flow.ipynb notebook for each RPi. The resulting masks are in a_processed_images/optical_flow folder

Once you have all processed images, run the mask finding pipeline:

7. Run the contour_finding.ipynb notebook notebook for each RPi and each method. The resulting masks are in a_found_masks/method/rpi folder.

   • Note: if you did not uncomment the commented code, you can run the next section

8. You can now run the a_xor_results/xor.ipynb notebook in order to get the XOR masks. The resulting masks are in a_xor_results/method folder.