Image Processing at Scale

Leverage cloud capacity and Activeeon elasticity

Cloud elasticity for Satellite Image Analysis


Request a Demo

We’ll send you maximum 1 email per month

Activeeon awarded with Innovation prize at Paris Open Source Summit 2018

For image processing and more

Image processing is more and more common in cases such as:

  • text recognition / optical character recognition (OCR)
  • object detection with deep learning
  • image segmentation with deep learning
  • image classification with machine learning
  • anomaly detection
  • change detection

The processing of images require specific resources such as GPU or a big enough scale. At the CNES, satellites images needs to be processed to standardized for further processing. 7PB of images needs to transformed. The cloud capacity is consequently adapted to process such scale.

The CNES is leveraging Activeeon scheduling solution to distribute the workload over an elastic resource pool in the cloud.

Other use cases such as registration plate recognition, crop growth analysis, …. are projects requiring large scale compute power.

A solution built to scale

Scale to 20,000 CPU cores and distribute your application at scale

Setup an elastic policy based on the load / queue and pay for what you need only

Implement an architecture agnostic to the cloud provider and stay in control between your on-prem needs and cloud needs.

Activeeon performance test charts
Request a demo Register and Try

Some use cases


Image processing for satellite images analysis,
registration plate identification, crop analysis and more.

3 photos for image processing

Why Activeeon

Distribution at scale and with precision

workload distributed on resource pool

Automatically select the appropriate resource for your workflow

Schedule and automate

automation icon

Build a strong basis for growth with template and standards. Give users advanced features with a user friendly interface.

Manage errors automatically

error management icon

Handle errors natively and select the rule to adopt from a dropdown menu.

Optimize

optimize resource consumption

Pay for what you need and nothing more

Any strategy: onprem, hybrid, multi-cloud

any cloud strategy

Embrace new resources with ease by adapting to your cloud strategy

End to end

dashboar from workload to resources

Control govern and monitor processes from the launch to the resource

Integrates with every process

Languages and Connectors

Java, Python, Javascript, Groovy, Powershell, Perl, Control-M, Dollar Universe, SLURM, PBS, LSF, SGE/OGE, Matlab, Scilab, Language R, Azure, Amazon Web Services (AWS), OpenStack, VMware, Docker, Talend, SAS, Hibernate, MySQL, Postgres, Oracle, MariaDB, …

For any situation

Move to open source solution for Batch Processing, Grid Computing, Hybrid Clouds, Java Workflow Engine, Job Scheduler, Open Source Workflow, Parallel Computing, Workflow Scheduling, Workload Automation

Screenshot from a workflow from ProActive Workflows and Scheduling

Testimonials

“Thanks to ProActive and additional optimizations we made along the way with Azure, the batch time which previously was taking us 18 hours, is now down to 5 hours, and by the end of the year, my hope and expectation is to go definitely below 4 hours and possibly less by having more grid hosts and more capacity.”

Guido Imperiale
Lead integration engineer at Legal and General

logo Capgemini, consulting company

“Activeeon is the only solution capable to Schedule any Big Data Analytics, mono-threaded, multi-threaded, multi-core, parallel and distributed.”

CapGemini Lead Engineer for Home Office

logo INRA, biotech company

“The ProActive product perfectly integrates into a hybrid multi-platform architecture. The usage of ProActive allowed us to have a trivialized approach to our processing infranstructure which includes Linux and Windows computers and GPU. This product helped us reduce operational costs by 10.”

Nicolas Pons
Head of InfoBioStat platform at MetaGenoPolis (INRA)