Voxelytics

Voxelytics offers an end-to-end AI solution to process and analyze your large-scale microscopy data. Optimized for Connectomics.
Reliable and highly optimized workflows to align datasets, segment cells, predict synapses, reconstruct neurons, generate connectomes, and more.

Discover the possibilities of Voxelytics

Neuron reconstruction made by scalable minds with Voxelytics. Raw SBEM data by Motta et al., Science 2019.

Features

Illustration

Automated microscopy data analysis

Voxelytics consists of powerful machine learning models built to reconstruct features from biological microscopy data. 
Train and run the models needed for your analysis and execute tasks such as neuron segmentation, synapse detection, segmentation of organelles, and more.

Illustration

Align and stitch EM datasets

Voxelytics supports volumetric imaging workflows from the beginning. Integrated routines for 3D alignment and in-plane stitching make it easy to turn a stack of individual image slices into a cohesive dataset for further analysis.
Dynamic mesh transformations and smart outlier detection enable smooth navigation through datasets.

Illustration

Straightforward evaluation of results

Run your predictions on an evaluation box and compare the results with your ground truth. 
Visualize and overlay your data in webKnossos to explore your reconstruction results.

Illustration

Advanced Deep Learning for neuroscience

Benefit from our investments in tuned AI model architectures and augmentation pipelines. Combine multiple tasks, such as membrane segmentation, type detection, and metric learning, in a single model. 
Easily use advanced features such as domain-adversarial training. Debug your models with rich evaluation and debug tools.

Illustration

Seamless data visualization with webKnossos

Visualize your results with the webKnossos. Interactively combine raw microscopy data with the Voxelytics predictions and derive insights. Use the handy annotation features to label landmarks, adjust image fidelity and navigate through peta-scale datasets.
Generate training data for new or fine-tuned models. Manually refine your segmentations with the powerful annotation and proof-reading capabilities of webKnossos.

Illustration

Full flexibility - without coding 

Take advantage of the endless possibilities of Voxelytics. Mix and match (pre-trained) models and tasks to compose your own AI and reconstruction workflows. 
Keep track of complex projects through the user-friendly interface and integrated results artifact management . Without a single line of code!

Illustration

HPC and Cloud Ready

Voxelytics is designed for working with peta-scale datasets on cloud services or your scientific high-performance computing cluster (HPC). Our tooling is optimized for throughput of very large datasets and heavily makes use of parallel computing and GPU acceleration.

    HPC and Cloud Ready
    Highly parallelized across many core and GPU systems
    Distributed computing with SLURM, PBS or other schedulers

Subscribe to our Newsletter

Receive regular updates when we publish new features for webKnossos or share updates on the Voxelytics tools and services.

Thank you!

We will contact you shortly.

Can't send form.

Please try again later.

How it works

1

Book an intro call

Discuss your research goals and data characteristics with us. Define the analysis tasks and arrange data access (ideally as webKnossos upload).

2

Receive a free segmentation sample for your data

Once we have access to your data, we will perform a segmentation on a subsample of your data (typically 1 GVx) with Voxelytics. Based on this, we can discuss the next steps and evaluate the need for re-training.

3

Optional retraining for your data

We have a large selection of pre-trained models for various types of EM images. However, sometimes it is required to retrain models for particular image characteristics. In that case, our annotators can generate the required ground truth and we will train custom models for optimal results.

4

Start the automated processing 

Roll out Voxelytics on your data. The processing pipeline includes stack alignment, neuron segmentation, neurite type detection, nuclei/somata/bloodvessel classification, synapse detection, and connectome assembly. 

5

 Polish your results in webKnossos

Visualize and evaluate the results in webKnossos. Use the advanced proofreading tools in webKnossos to correct any remaining errors on the objects you care about. Benefit from the collaboration features to speed up this process.

6

Work on your scientific analysis

Explore the results in webKnossos and use the available Python libraries for scientific analysis. Of course, you can download the data at any time!

Neuron reconstruction of Human cortex from volume EM. Raw data by Loomba et al. (Science 2022). Segmentation, connectome, and animation by scalable minds.

Case study

Loomba et al. submitted a comparative study of neural structures in 8 Mouse, Macaque, and Human datasets (SBEM, each 0.5-2TB). For that, a reliable, repeatable, highly-scalable solution for Connectome reconstruction across 3 different species with significant differences in cellular morphology was required.
The automated image alignment, neuron segmentation, and connectome reconstruction were carried out using Voxelytics.

We applied the default alignment workflows of Voxelytics on the respective dataset without manual tweaking. The out-of-the-box neuron segmentation and synapse detection workflows worked well for the mouse datasets. Any parameter tweaks could be made in a config file and consecutive Voxelyics runs would only execute the changed workflow tasks. To improve the reconstruction quality for the macaque and human tissue, we interactively re-trained the segmentation models with data labeled in webKnossos. We decided on a best-performing configuration after using the integrated evaluation methods and rolled that out to the remaining datasets. All Voxelytics workflows were executed highly parallelized on an HPC and results were instantly available in webKnossos for inspection.
Our declarative analysis approach, the repeatable workflows, and extensible task architecture allowed us to quickly iterate on the 8 datasets and derive insights in a manner of days. For scientists, Voxelytics moves the analysis burden from big data processing challenges to neurobiological analysis.

Use cases

Illustration

Connectome reconstruction for EM data

Run the Voxelytics ML pipeline to automatically reconstruct your neural connectivity matrix from EM images. Voxelytics ships with pre-trained AI models that work well for many EM datasets.
The pipeline consists of a sequence of tasks including:

    Alignment & Stitching of raw microscopy images
    Neuron reconstruction, including type prediction(dendrites, axons, spine heads...)
    Agglomeration heuristics to reduce split and merge errors
    Synapse detection and identification 
    Interactive connectome exploration with webKnossos
Request a free sample for your data
Illustration

Nuclei detection

Execute the Voxelytics model for nuclei detection on EM data. 
You will have your nuclei automatically segmented. Use the results such as the number of nuclei or their position to make meaningful scientific discoveries.

Illustration

Custom models and analysis

Create your own custom workflows to analyze your microscopy images. Voxelytics supports many image 2D and 3D imaging modalities (light and electron microscopy, Micro/Nano-CT, CT, MRI) and can be adapted to a wide range of image analysis tasks.It is optimized to work with large-scale volumetric EM inputs.
Train your custom models and quickly iterate to reconstruct, count, and measure the size of objects.

Illustration

Dense neuron segmentation of mouse layer 4 somatosensory cortex 

Full dense neuron instance segmentation using modified U-Nets and hierarchical agglomeration. Read blog article.

Explore in webKnossos
Illustration

Synapse, vesicle, and mitochondria detection in cortex

CNN-based segmentation of all synapses, vesicles, and mitochondria in preparation for synaptic connectivity mapping.

Explore in webKnossos
Illustration

Axon and dendrite classification

Integrate semantic segmentation of neuron subtypes (axon, dendrite, glia, etc) into the agglomeration to prevent merger error based on prior biological knowledge.

Explore in webKnossos

Pricing

Free sample

Illustration

Free

Receive a free Voxelytics sample reconstruction of your data to evaluate whether Voxelytics suits your needs. 

Get in touch with us to discuss data exchange and project goals.

Request sample

License

Illustration

Pricing upon request

Get a Voxelytics subscription to automatically process and analyze your data. The subscription includes a collection of pre-trained models, regular updates, an onboarding session, and support. 
Grant packages available

Get in touch

Services

Illustration

 

Hire us to support your data analysis projects. Our dedicated experts will collaborate closely with you to generate the reconstructions and analysis required for your project. 
Leave the big data challenges to us and focus on the scientific analysis.

Learn more

References

    • Year

    • Title

    • Authors

    • Publication / DOI

    • Year

    • Year

    • Title

    • Title

    • Authors

    • Authors

    • Publication / DOI

    • Publication / DOI

    • 2022

    • Connectomic comparison of mouse and human cortex

    • Sahil Loomba, Jakob Straehle, Vijayan Gangadharan, Natalie Heike, Abdelrahman Khalifa, Alessandro Motta, Niansheng Ju, Meike Sievers, Jens Gempt, Hanno S. Meyer & Moritz Helmstaedter

    • Science10.1126/science.abo0924

    • Year

    • 2022

    • Title

    • Connectomic comparison of mouse and human cortex

    • Authors

    • Sahil Loomba, Jakob Straehle, Vijayan Gangadharan, Natalie Heike, Abdelrahman Khalifa, Alessandro Motta, Niansheng Ju, Meike Sievers, Jens Gempt, Hanno S. Meyer & Moritz Helmstaedter

    • 2022

    • Functional and multiscale 3D structural investigation of brain tissue through correlative in vivo physiology, synchrotron micro-tomography and volume electron microscopy

    • Carles Bosch, Tobias Ackels, Alexandra Pacureanu, Y Zhang, Christopher J. Peddie, Manuel Berning, Norman Rzepka, Marie-Christine Zdora, Isabell Whiteley, Malte Storm, Anne Bonnin, Christoph Rau, Troy Margrie, Lucy Collinson & Andreas T. Schaefer

    • Nature10.1038/s41467-022-30199-6

    • Year

    • 2022

    • Title

    • Functional and multiscale 3D structural investigation of brain tissue through correlative in vivo physiology, synchrotron micro-tomography and volume electron microscopy

    • Authors

    • Carles Bosch, Tobias Ackels, Alexandra Pacureanu, Y Zhang, Christopher J. Peddie, Manuel Berning, Norman Rzepka, Marie-Christine Zdora, Isabell Whiteley, Malte Storm, Anne Bonnin, Christoph Rau, Troy Margrie, Lucy Collinson & Andreas T. Schaefer

    • 2021

    • Postnatal connectomic development of inhibition in mouse barrel cortex

    • Anjali Gour, Kevin M. Boergens, Natalie Heike, Yunfeng Hua, Philip Laserstein, Kun Song & Moritz Helmstaedter

    • Science10.1126/science.abb4534

    • Year

    • 2021

    • Title

    • Postnatal connectomic development of inhibition in mouse barrel cortex

    • Authors

    • Anjali Gour, Kevin M. Boergens, Natalie Heike, Yunfeng Hua, Philip Laserstein, Kun Song & Moritz Helmstaedter