First portable CellCanvas demo

Parent environment for supporting the Winter 2024 UX Evaluation

Fetch datasets for the UX evaluation in Winter 2024.

Run the UX evaluation itself

Export results of the UX evaluation.

A solution that predicts segmentation using a model for a Copick project and saves it as 'predictionsegmentation'.

Run CellCanvas

Run Surforama.

Apply a SwinUNETR model to a tomogram fetched using the Copick API to produce embeddings, and save them in a Zarr.

Automatically process all Zarr files within a specified directory structure using a SwinUNETR model.

Generate tomograms with polnet, and save them in a Zarr.

Fetches tomograms, tilt series, annotations, segmentations, and metadata from cryoet_data_portal and integrates them into the specified Copick project.

Convert a mrc to omezarr using mrc2omezarr.

Run CellCanvas with a copick project

Run CellCanvas with a copick project

Run CellCanvas with a copick project

Generates a DataFrame from picks and their corresponding embeddings for a single run and saves it.

Uses multithreading to compare median embeddings from multiple Zarr datasets to known class medians and identifies matches based on a configurable distance threshold.

A solution that paints picks from a Copick project into a segmentation layer in Zarr.

A solution that trains a Random Forest model using data from a Zarr zip store, filters runs with only one label, and performs cross-validation.

Compute multiscale basic features of a tomogram from a Copick run in chunks and save them using Copick's API.

A solution that processes all Copick runs and saves the resulting features and labels into a Zarr zip store.

A solution that optimizes a Random Forest model using Optuna, data from a Zarr zip store, and performs 10-fold cross-validation.

A solution that predicts binary segmentations for each label using models created by an optimization solution, and saves them separately.

A solution that extracts centroids from a multilabel segmentation using Copick and saves them as candidate picks.

A solution that trains a LightGBM model using data from a Zarr zip store, filters runs with only one label, and performs 10-fold cross-validation.

A solution that trains an XGBoost model using data from a Zarr zip store, filters runs with only one label, and performs 10-fold cross-validation.

A solution that optimizes an XGBoost model using Optuna, data from a Zarr zip store, and performs 10-fold cross-validation.

A solution that predicts segmentation using a model for a Copick project and saves it as 'predictionsegmentation'.

A solution that compares the picks from a reference user and session to a candidate user and session for all particle types, providing metrics like average distance, precision, recall, and F-beta score. Computes micro-averaged F-beta score across all runs if run_name is not provided.

Generates a file containing the median embeddings for each object type based on the picks in multiple runs, filtered by user IDs if provided.

Creates a new set of picks for a new session ID, containing the same locations but including the distance to the median embedding in the 'score' attribute.

Generate a Point Normal Surface (PNS) file from a PDB ID.

A solution that evaluates picks from a Copick project against a multilabel segmentation and computes metrics for each (user_id, session_id, object_name) pair for each run and across all runs.

A solution that splits datasets into training and test sets, ensuring distributions are preserved.

Train the SwinUnetr pixel embedding network using the provided script and dataset.

Run CoPick Live

A solution that places a grid of picks based on the radius of each pickable object, parameterized by a multiple of the particle radius, using tomogram shape.

Submit another album solution to Slurm as a job array by using the runs in a Copick project.

Train a UNet model to predict segmentation masks using MONAI from multiple runs with MLflow tracking.

A solution that opens a Copick project and displays the index using textual.

A solution that counts the number of voxels per label in a segmentation and saves the results as a CSV and HTML page.

A solution that installs napari-copick and launches the CopickPlugin.

A solution that uses the compare-picks album solution to evaluate all user_id and session_id pairs listed in a CSV file, creating JSON output files for each pair in a specified directory and submitting jobs to Slurm.

Train the SwinUnetr pixel embedding network using the provided script and dataset.

A solution that compares the rankings of candidates in the public and private test sets using various rank metrics.

Train the SwinUNETR pixel embedding network using the Copick dataset.

Train a 3D UNet network using the Copick dataset for segmentation.

A solution that exports multiple Copick runs' tomograms and picks into an HDF5 file.

Train a 3D UNet network using the Copick dataset for segmentation.

Train a 3D ResUNet network using the Copick dataset for segmentation.

Train a 3D Swin UNETR network using the Copick dataset for segmentation.

An album solution for copick monai morphospaces projects .

Generate segmentation masks using a trained UNet checkpoint on the Copick dataset.

An album solution for copick polnet projects .

Automatically generate segmentation masks for the Copick dataset. Install on a device with GPU access. Need cuda/12.1.1_530.30.02 and cudnn/8.9.7.29_cuda12 to compile.

Automatically generate Tensorboard event files and launch a visualizer for it. Currently suppot zarr and npy files.

A FastAPI server to manage Album solutions.

Backend for CellCanvas with Copick Config Support.

A solution that launches napari-cellcanvas with optional config fetching.

Fetches a Copick config from a FastAPI server and writes it to a file.

A solution that opens a Copick project and displays the index using Panel.

Train the Binary SwinUNETR pixel embedding network using the Copick dataset.

Generate a Point Normal Surface (PNS) file from a PDB ID.

Compute multiscale basic features of a tomogram from a Copick run in chunks and save them using Copick's API.

A solution that creates mock annotations based on multilabel segmentation, trains XGBoost models in steps, and generates predictions.

Evaluates segmentation models from the mock-annotation solution on Copick data, generating metrics like IoU and F1, and saves the predicted segmentation into the Copick project.

A solution that creates mock annotations based on multilabel segmentation, trains a PyTorch segmentation model in steps, and generates predictions.

Backend for CellCanvas with Copick Config Support.

A solution that processes Copick runs, filters runs with only one label, and trains an XGBoost model with 10-fold cross-validation.