Data entry ops · Production

NASA JPL uses Labelbox video annotation to build life-detection ML algorithm 5x faster

The problem

The small MLIA team at NASA JPL needed to label large, diverse microscopy video datasets to train their life-detection ML model, but building and maintaining an in-house labeling solution was burdensome — PhD researchers were painstakingly writing Java GUIs and manually sorting data instead of doing ML work.

Workflow diagram · grounded in source
1
Microscopy video collected
trigger
“The team quickly uploaded video datasets collected from Digital Holographic Microscopes”
2
Upload to Labelbox
integration
“We just had to upload our data, customize the editor to our exact requirements, and go”
3
Workforce video annotation
human_review
“the labelers were easily able to track various objects' movements over time and determine if it might be a live organism”
4
Scientist model review
human_review
“scientists are integrated into the ML pipeline. The MLIA team coordinates with them to review model outputs and verify whether their model tracks the right objects, folding their expertise into their iterative cycles throughout the process”
5
ML life-motion detection
ai_action
“build an ML model that identifies microscopy videos most likely to contain signs of life-like motion, captures short clips of the potentially living creatures, and prioritizes them to downlink back to Earth along with an explanation of i…”
6
Prioritize clips for downlink
output
“prioritizes them to downlink back to Earth along with an explanation of its decisions”
Reported outcome

Using Labelbox, the MLIA team achieved 5x the speed for editor customization and video annotation, completing setup in a single day against a budgeted week, while Boost delivered significant improvements to their model.

Reported metrics
Annotation and editor customization speed5x
Setup time vs. budgetbudgeted a week, completed in a day
Data return ratio vs. data gathered<0.01%
Model improvementsignificant improvements in our model
Reported stack
LabelboxLabelbox BoostLabelbox WorkforceDigital Holographic Microscopesdecision treesSVMs
Source
https://labelbox.com/customers/nasa-jpl
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Using Labelbox, the MLIA team achieved 5x the speed for editor customization and video annotation, completing setup in a single day against a budgeted week, while Boost delivered significant improvements to their model.

What tools did this team use?

Labelbox, Labelbox Boost, Labelbox Workforce, Digital Holographic Microscopes, decision trees, SVMs.

What results were reported?

Annotation and editor customization speed: 5x; Setup time vs. budget: budgeted a week, completed in a day; Data return ratio vs. data gathered: <0.01%; Model improvement: significant improvements in our model (source-reported, not independently verified).

How is this data entry ops AI workflow structured?

Microscopy video collected → Upload to Labelbox → Workforce video annotation → Scientist model review → ML life-motion detection → Prioritize clips for downlink.

Related data entry ops cases