Applications

Enable Smart on-device AINR 

Our AI noise reduction is the first high performance model suitable for face-to-face conversations in noisy environments. It can be configured for ultra-low-power or ultra-low-latency, and in single or multi-microphone configurations. This technology can be deployed in earbuds, hearing aids, and headsets to significantly improve speech intelligibility in noisy environments while maintaining exceptionally long battery life. It can also be used as a front end to speech recognition systems to improve accuracy by 50%!

Future Hearables

In the near future, hearables will be equipped with PPG, FNRIS, and EEG sensors, providing troves of data that can be processed by emerging algorithms running on the SPU to provide real-time insights about a user’s health, stress, cognitive load, attention, and overall well being without sending constant streams of data to external devices or the cloud.

The hearable is no longer just an audio streaming device. It’s a secure and compact hub for your entertainment, senses, health, and productivity!

Enable the convenience of voice control everywhere

Femtosense’s smart device algorithm suite allows virtually any device to be transformed into a multi-functioning smart device. With wake word detection, keyword spotting, and spoken language understanding models, we enable an efficient voice interface pipeline for even the most cost sensitive devices, delivering the comfort and accessibility of voice control anywhere. 

Sparsity makes it possible

Thanks to sparsity, we are able to offer more AI applications in the same size and cost footprint. This can transform devices like remote controls, smoke alarms, and cameras into baby monitors, security systems, and communication devices without significant modification or effect on product margins.

More value in everyday products offers manufacturers a significant competitive advantage, and users peace of mind and convenience. Everyone wins!

Protect your most precious assets

Anomaly detection algorithms allow vibration sensors and microphones to detect failures in all types of machinery. Always-on preventative maintenance systems that constantly stream data to a central hub for analysis can be complex to deploy and burdensome to upkeep. 

Sparsity enables easy scalability 

With SPU technology embedded, IMU sensors and microphones can become more intelligent, running failure detection algorithms on-device and only instructing host processors to send data when something is detected. By exploiting sparsity, anomaly detection algorithms can be heavily compressed and run for longer periods of time without changing sensor batteries.

IMU sensor and microphone manufacturers can offer more value to customers with differentiated products. Sensing system designers can improve the battery life and robustness of their systems. Heavy machinery, vehicles, HVAC systems, and entire manufacturing lines can be monitored and protected from failures affecting production and functionality.