Artificial intelligence (AI) hardware vendor Syntiant Corp. has debuted a neural decision processor (NDP), the first chip from the company to provide vision processing for next-generation AI applications.
The NDP200 chip is packaged with the Syntiant Core 2 and can run multiple applications simultaneously with low power consumption including AI vision for person detection, object classification, motion tracking and occupancy monitoring.
The chip can also perform processing on-device including multi-sensor fusion, voice command recognition, acoustic event detection, infrared detection, anomaly and tamper detection, and other audio, motion and pressure sensing recognition.
The always-on neural processing could be used for all types of consumer and industrial battery-powered devices such as IP and security cameras, smartphones, tablets, smart speakers, smart displays and smart home applications and security devices.
Features include:
- Hardware acceleration more than 6.4 GOP/sec.
- Support for neural networks, recurrent neural networks, fully connected networks and can process multiple heterogenous networks concurrently.
- Support for more than 7 million parameters.
- Embedded user programmable HiFi 3 DSP.
- Embedded Arm Cortex-M0 processor.
- 11 wire direct image interface.
- SPI and I2C for multi-modal sensor fusion.
- Dual PDM microphone interface.
- 12S serial interface with PCM.
Syntiant Core 2 is a deep learning inference engine that moves larger neural networks into always-on domains with the capability to generate shared embeddings, run ensembles and other neural architectures concurrently or in cascades. The Core 2 engine provides full control of multiple independent networks while enabling machine learning deployment without any intermediary compilers.
