Open Science Portal
We are fully committed to open science: open-source software, public datasets and benchmarks built for AI research, and open-access publications. Open science makes impactful science. Below you find the tools and data released with our papers and earlier open-science work by our group members.
Open-source software
Web app to search, organize, aggregate and share image data via metadata (web app, our portal scripts)
Code
Deep-learning-based compression for microscopy images, with trained checkpoints
Code
Cell segmentation and tracking pipeline for the ComplexEye array microscope
Code
napari-powered tool for mass spectrometry imaging data analysis
Code
Masked diffusion as a self-supervised representation learner for segmentation
Code
Rethinking brain MRI anomaly detection from an image quality assessment perspective
Code
Data-centric active learning for biomedical image instance segmentation
Code
Code for studying redundancy and forgetting in biomedical segmentation with Cellpose
Code
Code for “On the risk of manual annotations in 3D confocal microscopy image segmentation”
Code
Our solution to the LightMyCells challenge for label-free organelle prediction
Code
Analysis code for nuclear growth trajectories in hiPS cell colonies
CodeOpen datasets
Benchmark of end-to-end bioimage analysis tasks for evaluating AI agents (code, project page)
Data
About 8,000 brightfield neutrophil images labelled for quality issues (BioImage Archive)
Data
Raw time-lapse movies of immune cell migration from the ComplexEye array microscope
Data
Brightfield neutrophil videos from ComplexEye, used for cell tracking and FlowRoI compression
Data
Large-scale collection of 3D medical shapes for computer vision (web interface and Python API)
Website
Dataset used to train the 3D mitotic classification
DataEarlier work by our group members
This is the ML part of Allen Cell and Structure Segmenter software
Code
This is the classic part of Allen Cell and Structure Segmenter software
Code