FB pixel

Aryballe gets more funding for edge devices that can ‘smell’: Will digital olfaction finally pass the sniff test?

Categories Biometric R&D  |  Biometrics News  |  Trade Notes
Aryballe gets more funding for edge devices that can ‘smell’: Will digital olfaction finally pass the sniff test?

Grenoble, France-based Aryballe has a different take on edge computing devices – their sensors can smell, in a manner of speaking. Now, the company has raised €7M in additional funding to continue to develop its digital olfaction technology. Among the potential uses for the technology: identifying people based on smell as a biomarker and detecting medical conditions such as COVID-19 infection.

Digital olfaction is a field of computing that seeks to find ways to detect and classify scents in our environment. The idea is based on the process by which a human nose can detect molecules; cells in the nose send electrical signals to the brain, where groups of interconnected neurons turn those signals into the sensation of smell. Similarly, digital olfaction takes signals from sensors and seeks to classify electronic signals according to learned patterns in data. In other words, sensors are ‘trained’ to tell different scents apart based on chemical composition and other characteristics.

Currently, scientists are conducting experiments to see if dogs can be trained to detect COVID-19; similar experiments with the detection of cancer have shown promise. It’s not a stretch to suggest that digital sensors could eventually be trained in the same way, though Aryballe’s initial applications are in areas such as food processing (detecting the quality of a vegetable or fruit) and automotive (detecting smells that indicate imminent failure of a part).

Aryballe says it has combined biochemical sensors, advanced optics, and machine learning in a single solution to collect, display and analyze odor data. The company has been working on the technology since 2014; last year the company noted that it has now secured 17 patents for its technology.

The funding comes from new investors Samsung Venture Investment Corp (Samsung Ventures) and Seb Alliance (Groupe SEB’s corporate venture arm) as well as existing investors Innovacom, Cemag Invest, Asahi Kasei, and HCVC. The round brings the company’s total funding to date to €17M.

The field of digital olfaction is not new, but recent advances in chipmaking and AI/ML processing have helped spur a resurgence of activity. Airbus has partnered with Koniku, a startup in the biotech sensor field. The focus of the project was initially around security issues such as bomb detection but has expanded to include COVID-19 detection.

Similarly, Intel detailed its research around the application of neuromorphic computing chips such as its Loihi processor for digital olfaction.

Nvidia gains another imaging partner for edge

The widespread use of digital olfaction for biometric identification is probably still years off in the future. Digital imaging solutions that leverage the growing power of edge computing systems are going to be a key tool for biometric systems for the foreseeable future.

Nvidia has been helping build that wave with products such as the Nvidia Jetson family of edge devices with GPU processors. The company is frequently adding technology partners to its Partner Network program, which helps bring new applications that use its platform to market. NexOptic Technology Corp. recently announced that it has joined the program, saying that it is offering its ALIIS (All Light Intelligent Imaging Solutions) technology on Nvidia’s platform. ALLIS is a set of image-enhancing algorithms that use inferencing to provide improved resolution and reduced motion blur, among other stated improvements.

Related Posts

Article Topics

 |   |   |   |   |   |   |   | 

Latest Biometrics News

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Most Read This Week

Featured Company

Biometrics Research

Biometrics White Papers

Biometrics Events

Explaining Biometrics