Imagimob is featured in Dagens Industri, the leading business newspaper in Sweden. You can read the article here (article in Swedish).
Here is the article translated:
In 2020, Imagimob launched its SaaS platform Imagimob AI, which enables fast and easy development of innovative Edge AI applications for small chips that have limited computing power and memory capacity. Imagimob AI streamlines the journey from idea to commercialization, which leads to increased productivity and faster time-to-market.
Digitization has paved the way for more and more products with built-in intelligence that together create increasingly smarter, connected ecosystems that help us make our lives accessible, easier, more efficient and optimized. Imagimob specializes in sensor technology and AI, and they develop and sell their own SaaS service (Software as a Service). The SaaS service assists with the technology that enables product development companies to quickly, easily and cost-effectively further develop and provide their products with intelligence - or create completely new products and services - with AI running on small chips, so-called Edge AI or tinyML.
- Building Edge AI applications for small chips is a complex process, and it requires both a competent multidisciplinary team and advanced software support, says Anders Hardebring, co-founder and CEO of Imagimob AB.
- For a company to implement Edge AI in an existing or a new product, a number of different skills are required. You need a domain expert who understands the product area, and who can define what you want to detect, prevent or streamline, for example. You need a data analyst who can understand and analyze the data generated and build the right type of AI models. And you need someone who can optimize and integrate the AI model on the chip itself.
Alexander Samuelsson, co-founder and CTO adds:
- Instead of the company employing or hiring all these people as consultants, who can be difficult to find, we can offer them a flexible SaaS solution that reduces costs and streamlines the journey from idea to commercialization. The customer is offered greater control over the entire development chain. At the same time, our business also becomes scalable, as we sell the service itself, which makes the business model advantageous for both parties.
Together with Imagimob's expertise in Edge AI, the SaaS service Imagimob AI opens up fo a number of different applications in a broad and rapidly growing market.
- Imagimob AI is basically a horizontal service that supports technology for deep learning and all types of time series data. But Imagimob has chosen to delve into different vertical solutions, and the strategy is to become a world leader in each chosen vertical.
One such vertical solution is gesture control, which in conjunction with the corona pandemic is a highly topical functionality for combating the spread of infection through the exclusion of physical contact. In collaboration with the radar sensor company Acconeer and the ODM supplier OSM Group, Imagimob has developed gesture-controlled in-ear headphones, which were shown at a large trade fair in early March.
- Instead of taking out the mobile phone, pressing or turning the device to change song, answering incoming calls or raising and lowering the volume, it is enough to bring your finger up to your ear and make one of four possible, learned gestures.
But the potential for learning a larger repertoire of gestures is great. To recognize gestures with high accuracy, the AI must be trained with data from many different users. This means that the gestures are perceived as very robust, despite the fact that hand movements by nature always differ somewhat from person to person.
- The technology is of course not only limited to headphones, but is applicable to many different types of electronic devices. Wherever gesture control and non-contact user interfaces can in various ways facilitate and optimize the use of products and systems, for example in healthcare, in the automotive industry and in the manufacturing.
Another area that is now generating a lot of customer inquiries is different types of audio applications.
- Many customers are looking for AI that can recognize speech and different types of sound. For example, it can be used to recognize different sounds in an urban environment or in an industrial environment. It can also be about offering other types of sensory signals to inform, notify or alert people with hearing loss, to facilitate navigation in different environments, or to improve the interaction with different devices. The market is young and so far we have only scratched the surface of a very large area, where very exciting things will happen in the next few years.
Trying to put an AI model in a real-world scenario can be da...