Training and inference discussion for active learning round 2

[rohanbanerjee]

Continuation from the previous round of training: #35

What is the round 2 model

The model which was fine-tuned on the manually corrected segmentations as per the QCs mentioned in #35 is the round 2 model. A total of 26 images were added in the training of this model since we fine-tuned the previously trained round 1 model.

A list of subjects used for the fine-tuning is below: finetuning.yml

The config (containing preprocessing, hyperparameters) for nnUNetv2 training is: plans.json

After the training was completed, I ran inference on the rest of the images whose segmentations have to be included in the consequent rounds of training (226 images), below is the QC. 40 subjects from these images will be chosen and included in the round 3 of training:

qc_round2_inference.zip

The steps to reproduce the above QC results (/run inference) are the following:

1. Clone this repos
2. cd fmri-segmentation
3. Download the model weights (the whole folder) from the link: https://drive.google.com/drive/folders/1WSn-15wGWz6i2_aZeQTwKls2sZ6dpfHf?usp=share_link
4. Install dependencies

pip install -r run_nnunet_inference_requirements.txt

5. Run the command:

python run_nnunet_inference.py --path-dataset <PATH TO FOLDER CONTAINING IMAGES, SUFFIXED WITH _0000> --path-out <PATH TO O/P FOLDER> --path-model <PATH TO DOWNLOADED WEIGHTS FOLDER>

Next steps:

• Add inference QCs in this comment
• Choose and correct 40 segmentations and include it in the next training round along with the ones that have good segmentation after inference.
• Get metric by running evaluation on the held-out test set (Creation of a held-out test data for active learning training phase validation #33)

[rohanbanerjee]

The list of 40 subjects chosen for manual corrections is below:
qc_fail.yml.zip

The QC report for the segmentations manually corrected by me for the above 40 subjects is below:

qc_round_2_corrected.zip

@jcohenadad , I would like your inputs for the above manually corrected images. I will use these images for the next round of training.

CC: @MerveKaptan

[rohanbanerjee]

Julien: Redo KCL subjects

[jcohenadad]

here you go: qc_flags.json

two issues only:

[rohanbanerjee]

Closing the issue since the round 2 training was successfully completed (including running inference and manual correction)