Features

Visualize, share and annotate your large 3D images online

WEBKNOSSOS is open source and has been published in Nature Methods.

Fast browsing of massive datasets–from anywhere

Specialized storage, streaming and rendering technology enables fast browsing speeds to your work computer, home-office laptop and tablet on-the-go. There is no need to install anything beyond the browser. This makes it easy to onboard collaborators.
WEBKNOSSOS is optimized for datasets ranging from gigabytes to petabytes in size. We developed a custom chunked storage infrastructure, streaming-optimized backend software and GPU-accelerated rendering to enable fast data access on most devices, including laptops and iPads.

Volume annotation

Create volume annotations (i.e. segmentations) with manual brush tools. Use the interpolation feature to speed up 3D annotations. Download the annotations to train a machine learning model or for visualization purposes.
Visualize segmented objects as meshes through the integrated mesh generation. Explore dense segmentations with colored and patterned maps.

Example: Manually segmenting nuclei in mouse cortex, EM data from Motta et al. 2019

Unique skeleton annotation

Create skeleton (i.e. line-segments) annotations of neurons, measure their path lengths and organize these skeletons in hierarchical groups. Attach comments to nodes in order to mark interesting locations your data.
Try out the unique Flight Mode for high-speed tracing of axons or dendrites. Trained annotator crowds achieve tracing speeds of 1.5 ± 0.6 mm/h for axons and 2.1 ± 0.9 mm/h for dendrites in 3D electron microscopy data. Read more in the WEBKNOSSOS paper.

Get started and upload your first dataset for free

 Collaboration and secure data sharing

Invite your colleagues and collaborators to WEBKNOSSOS to share your data with them. Because your data is in the cloud, every dataset and annotation has a link that you can share. 
No worries, by default your data is only privately accessible by you and collaborators you choose. Use protected links to share data with reviewers or outside collaborators.

Mesh generation and visualization

Explore your data in 3D with automatically generated meshes. WEBKNOSSOS generates meshes from volume annotations or pre-computed segmentations.
Use the screenshot feature to create figures or export the meshes in your favorite 3D rendering software (e.g., Blender, Amira, Cinema4D).

Visualization of machine learning predictions

Add prediction maps and color-coded segmentation layers to your dataset for debugging your work. Create and apply mappings (combining multiple segments into one) to evaluate agglomeration strategies.
Visually inspect prediction errors with the context of the raw data and other channels. Use the histogram to manually select classification thresholds.

Proofreading of automated segmentations

Correct split and merge errors with the proofreading tools. Visualize your segmented objects as 3D meshes and proofread directly in the 3D viewport. Highlight the segments forming an agglomerate by hovering to identify potential edits. Use Merger Mode to fix over-segmentations. Use the skeleton tools to quickly connect segments.Use the volume annotation features to correct any errors in your automated segmentation.

Public data hosting for the community

Tell your story with data. Link directly from a figure in your publication to that location in WEBKNOSSOS. Readers will be able to explore your annotations and understand the context of your findings. 
WEBKNOSSOS is an excellent platform for publishing large datasets including segmentations and training data. Viewers can freely browse through your data and build upon it. 

Data management features

Use WEBKNOSSOS as a centralized storage for your large-scale datasets. Stop scattering data around endless hard-drives or storage servers. WEBKNOSSOS easily scales to petabytes of data.
Fine-grained user access levels and subteams help you manage datasets securely within your organization.

AI analysis features

When annotating manually, use the AI-based quick-select tool to segment cells in 3D with a single click.
For automated segmentation, run our AI neuron model to segment an entire region of your data. 
The automated analysis features of WEBKNOSSOS use machine learning classifiers for dataset segmentations. These features are still in beta phase, every feedback is welcome. 

Looking for Automated Image Analysis?

Check out our image analysis services. Reconstruct rich information from microscopy images with machine learning-based tools. Automate tasks such as image alignment, registration, dense segmentation, and object detection.

Annotation task and project management

Set up tasks for your annotators and use auto-assignment to distribute the work. Works for both skeleton and volume annotation tasks. 
Keep track of the progress with monitoring reports. Download all created annotations in a merged file for further processing.
If you don't have annotators, you can hire our annotation services directly through WEBKNOSSOS.

Auto-save and versioning

WEBKNOSSOS auto-saves annotations every 30s and keeps a versioned history of every action in all annotations. Correcting a mistake is just one "undo" away.
Annotation data on WEBKNOSSOS is externally backed up daily. For paid plans, there is an option to back up datasets as well.

Scripting

Implement your own annotation tools with the flexible frontend API. Write JavaScript code to add shortcuts, query data, write comments, create mappings and much more. Learn more in the frontend documentation.
Use our free Python library to download annotations, upload segmentations or create tasks programmatically.

Interoperable data formats

Download your data and annotations from WEBKNOSSOS to work with them in other tools. WEBKNOSSOS supports standard formats (e.g. TIFF, STL, N5/ZARR, CSV) for exports.
Work with the WEBKNOSSOS file formats in Python or MATLAB, with our open-source libraries. Learn more in the user documentation on tooling.

Supported image modalities

WEBKNOSSOS works with many sorts of 2D and 3D image modalities including multi-channel data.

    Scanning electron microscopy (SEM)
    Transmission electron microscopy (TEM)
    Serial block-face scanning electron microscopy (SBEM)
    Focused ion beam scanning electron microscopy (FIB-SEM)
    Serial section electron microscopy (ssSEM, S3EM, ssTEM)
    X-ray tomography (CT) and Micro-CT (µCT)
    Magnetic resonance imaging (MRI)
    Fluorescence microscopy

Examples (from left): Fluorescence microscopy from Drawitsch et al. 2018, Synchrotron X-Ray Tomography from Kuan et al. 2020 and MRI from Lüsebrink et al. 2017

Illustration

Looking for Neuron and Connectome reconstruction?

Check out our image analysis services for neuron and connectome reconstructions from volume EM data. Automate tasks such as image alignment, registration, dense segmentation, and object detection.