Virtual interactive event provides a first-look at the TENSAI Platform and People Detection and People Counting demo; allows attendees to text or chat 1-1 with Eta Compute experts.
WESTLAKE VILLAGE, Calif., September 1, 2020 – Eta Compute Inc., a company dedicated to delivering machine learning to low power IoT and edge devices using its revolutionary TENSAI® Platform, announced it will exhibit at the 2020 Embedded Vision Summit, the premier event for innovators adding computer vision and AI to products.
This year’s Summit will be 100% online, allowing attendees to watch presentations, engage with speakers, visit the virtual exhibit hall, view live product demos on a seamless virtual platform. The event takes place September 15, 17, 22, and 24, 2020, from 9:00 a.m. to 2:00 p.m. PT,
Eta Compute’s Interactive Virtual Booth will feature its TENSAI® Platform that enables ultra-low power machine vision for edge devices. The TENSAI Platform includes the recently launched TENSAI Flow software and neural network compiler, which enables seamless development for machine learning applications in IoT and low power edge devices.
Leveraging TENSAI Flow and its neural network compiler, Eta Compute will demonstrate three uniquely optimized vision algorithms using low power CMOS image sensors:
- CIFAR10 object detection algorithm: Compared to direct implementation on a competitive device Eta Compute’s CIFAR10 neural network improves energy per inference to a staggering 200x factor.
- Person detection algorithm: Detect people in 3mW and 0.5 inference per second.
- People counting algorithm: Count people in less than 5mw and 1 inference per second.
All three neural networks are optimized for efficiency and memory usage and can run on an ultra-low power TENSAI Neural Sensor Processor from Eta Compute.
The people detection and people counting networks are typically used in space utilization and social distancing monitoring in office buildings and retail stores. This type