Evaluation Kit for NeuroVoice AI voice detection
POLYN Technology has launched an evaluation kit for its Neuromorphic Analog Signal Processing (NASP) technology and products for AI voice detection.
The AI voice detection evaluation kit from POLYN includes the product’s digital twin, which is available online, and evaluation boards with a dedicated user interface.
NeuroVoice is a Neuromorphic Front End (NFE) product family for voice detection, voice control, and clear communication in noisy environments. It is based on the NASP analog AI chip that has a power consumption between 35-100 microwatts and low inference latency at about a few microseconds.
NeuroVoice Smart Microphone and Smart Voice Control with Word Pattern Extraction evaluation boards contain a digital simulation of the neural core that mirrors the IC version. They operate with a live audio signal, enabling users to evaluate the neural cores live under different conditions. In addition, they provide the SDK for word classifier uptrain to adjust to special KWS requirements.
Neuromorphic front-end for tire monitoring
The evaluation board helps potential customers start the design process for intelligent voice management systems, even before chip availability.
The online digital twin – D-MVP Portal – is a web application designed to simulate the operation of a neural network so that the NASP model can be tested and adjusted to customer needs. The NASP model has already been transferred to Netlist but still runs on a digital processor.
The test procedure involves uploading various audio files, processing them by VAD, VE, and word recognition (SVC) neural cores, and generating results ready for playback and analysis.
“The NeuroVoice Evaluation Kit results from thoroughly examining customers’ needs during a roadshow with detailed discussions that POLYN started this spring. Those discussions continue with other interested parties in the year’s second half, and vendors are invited to evaluate it,” said Eugene Zetserov, POLYN’s Vice President of Marketing and Business Development. “The SDK (the Classifier reference code) allows potential customers to get closer to new product development before incurring chip manufacturing costs.”
“Adoption of Edge AI is still slow because of a few barriers, including complex integration of novel hardware and software that is vital for the successful implementation,” he said. “The Evaluation Kit addresses this concern.”
If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :
