The circuit is optimized for entry-level driver monitoring for processing eye gaze and eye-tracking algorithms. The ASIC plus an image sensor would consume less than 1W combined.
The OAX8000 uses a stacked-die architecture to provide 1Gbyte of DDR3 SDRAM. The neural processing unit (NPU) and image signal processor (ISP) provide dedicated processing speeds of up to 1.1 trillion operations per second for eye gaze and eye tracking algorithms.
The OAX8000’s on-chip NPU is supported by the TensorFlow, Caffe, MXNet and ONNX tool chains. Additionally, this ASIC embeds quad Arm Cortex-A5 CPU cores with Neon technology for accelerated video encoding/decoding and on-chip video analytics algorithms, along with hardware for image processing, video encoding and RGB/IR processing. The integrated video encoder accepts up to 5 megapixel captures from OmniVision’s automotive image sensors, and outputs up to 2K resolution video at 30 frames per second (fps).
Other applications include processing occupant detection algorithms and providing alerts when objects are left behind in the vehicle. Additionally, this ASIC can be used in automotive video security systems to perform functions such as FaceID, as well as preset driver-comfort settings that are activated when the DMS first scans the driver’s face.
Samples of the new OAX8000 ASIC are available now in a BGA196 package. It is AEC-Q100 Grade 2 certified for automotive applications.
Related links and articles: