AI
AISIN CORPORATION

AISIN CORPORATION

To run more than a few dozen AI models — facial landmarks, detection of worn items, gaze detection and more — on an inexpensive CPU, for the driver monitoring system in Toyota’s advanced driver assistance system “Advanced Drive.”

The challenge Idein solves

The request came when the hardware and the Arm-core processor had already been decided. With the processor fixed, the only remaining option was to make the models faster on the side being deployed. Idein contributed optimizing compiler technology that accelerates deep learning inference, along with research and development on lightweight models, and had it running in three to four months. That collaboration began in 2017 and has since extended to a low-speed autonomous-driving path planner and multimodal agents.

What we did

The hardware and the Arm-core processor were already fixed, and we turned to Idein for help getting these AI models to run at all within that. They delivered results in about three to four months.

— Susumu Osuga, Chief Engineer, Advanced Development Department, AISIN CORPORATION

  • The collaboration began in 2017, on the driver monitoring system for Toyota’s advanced driver assistance system “Advanced Drive”
  • More than a few dozen AI models — facial landmarks, detection of worn items, gaze detection — deployed on an inexpensive CPU, starting from fixed hardware and processor
  • A capital and business alliance followed in February 2018 (5% equity stake). AISIN was the first customer company to invest in Idein
  • Several AISIN employees now work at Idein on secondment for development experience
  • Low-speed autonomous-driving path planner: developing a path planner for autonomous driving in parking lots and private land, where — unlike public roads — traffic lights, signs and lane markings are not standardized. This was Idein’s first time building a deep-learning-based path planner, so the team supported the project from the literature survey onward, handling not just the algorithm but also building out the development environment
  • Multimodal agent “Saya”: developing AI agent technology that converses naturally with people. Everything runs on a single palm-sized computer, and by reading not just speech but also face orientation and mouth movement through the camera, the team built “multimodal utterance-detection technology” that keeps conversations from breaking down even when many people speak up one after another
  • Related releases: AISIN SEIKI Co., Ltd. and Idein Inc. enter a capital and business alliance (Idein, PR TIMES) / Running as-is on Raspberry Pi as the advantage: edge AI company Idein deepens its collaboration with AISIN (MONOist)

The next site is your site

Don’t see a close match among our case studies? Tell us about your problem anyway.