Startup Jensorter has created an AI-powered fish egg sorter that quickly tells a healthy versus an unhealthy fish egg. The sorting machine uses an Nvidia Jetson Nano module, according to Nvidia.
A Jetson module is a small form-factor, high-performance computer that can be embedded with sensors, AI software and other services.
Jensorter’s device detects heathy vs. unhealthy eggs and also egg size and fertility status.
Customers in Alaska, the Pacific Northwest and Russia use the machines to separate viable eggs from unhealthy ones. These fisheries normally collect millions of fish eggs regularly to protect them from predators to increase yield. Once collected, fishery operators need to separate healthy eggs from those infected with parasites that put the healthy ones at risk
The Jensorter machines use characteristics like color to find a healthy egg and detect whether it is fertilized, and can do so at speeds of 30 milliseconds per egg. The fish egg sorters are much higher accuracy than other methods, said Curt Edmondson, CTO of the Oregon-based startup.
Jensorter’s Model JH device can sort nearly 200,000 eggs per hour, an obvious replacement for sorting by hand.
Edmondson told Nvidia that the use of AI to capture and process images of eggs could transform the industry if hatcheries would share images in a database to identify patterns of egg characteristics that lead to healthy eggs. Such an approach would help propagate species on the decline in some areas of the world, including salmon and trout.
This spring, the Oregon Hatchery Research Center plans to use Jensorter machines to beta test at hatcheries. Jensorter has posted a tutorial from 2015 on how its equipment works. Fish eggs are examined individually by getting lodged into holes on a spinning disk.
RELATED: Nvidia planning to say bye-bye to troubled Arm deal, sources say