Case Study: NASA’s Jet Propulsion Laboratory speeds up expert video labeling and improves life-detection modeling with Labelbox

NASA’s Jet Propulsion Laboratory sets up video evaluation 5x faster with Labelbox

NASA's Jet Propulsion Laboratory (JPL) faced the challenge of building a machine learning model to detect signs of life in video footage from potential ocean worlds in our solar system. Their small team needed to process a large, diverse set of water sample videos but found that building and maintaining an in-house data labeling solution was a major obstacle. They turned to Labelbox's video evaluation solution to efficiently gather the expert feedback required to train their model.

Labelbox provided a platform where contributors could easily track object movement over time and judge if motion was life-like. Using Labelbox, JPL customized the editor to their exact requirements and completed their most manual-intensive tasks in a day, a process they had budgeted a week for, setting up about five times faster than expected. The high-quality expert feedback delivered through Labelbox's managed services drove significant improvements in their model.

NASA’s Jet Propulsion Laboratory (jpl)