Beyond labels: Tellspec case study

Tellspec has designed a portable food testing solution that will help consumers find out what’s in their food even if there’s no label.

The testing technology scans food extremely quickly and non-invasively using near-infrared (NIR) spectroscopy – a light-based technique for identifying materials based on reflected wavelengths. The food is first scanned by the portable sensor which shines a beam of light into the item and then measures the wavelengths reflected back after some of the energy has been absorbed by the food; this produces a spectrum. The data is sent via bluetooth to the app on the user’s smartphone. Once the smartphone receives the digital spectrum of the food, it transmits it to the Tellspec analysis engine which processes the data, compares it to reference spectra and runs a complex set of algorithms on it that use machine learning and bioinformatics techniques. Then the results are sent back to the Tellspec app on the user’s smartphone, all within a few seconds.

This short film provides a visual demonstration of how the technology works.

The combination of machine learning, bioinformatics and traditional spectroscopy gives Tellspec the ability to extract a lot of information from a spectrum, providing users with specific details about the chemical nature of the food being analysed. The algorithms can count calories, identify the amounts of dairy, fat, proteins, fibre, carbohydrates, sugars, gluten and soy, and calculate the glycaemic index within seconds, helping users to select foods that meet their dietary requirements.

The technology to break down food to analyse its composition has been available for a long time – but for science labs, usually in bulky and expensive instruments and not for consumers. Tellspec developed a consumer friendly, pocket-sized food sensor, small enough to fit into the palm of a hand that can be potentially implemented into mobile phones. In addition, the sensor can be operated without any knowledge in the field of chemistry and physics and its aimed at consumers who want to make educated, science-based food choices.

However, caution should be applied with the instruments. The primary concern expressed by the experts in the field (NIR spectroscopy community) was that miniaturised equipment still requires comprehensive validation studies performed in well-equipped laboratories [1]. The opinion of other experts is that the ‘cloud-derived’ data, advertised by several direct-to-consumer companies is not realistic, as their food-scanners are relatively simple and therefore cannot achieve high accuracy [2].

Nevertheless, Tellspec food sensor it’s used by companies to ensure process and product quality control, including a tool that measures the sweetness or acidity of fruit and vegetables (‘FruitQC’), allowing them to scan their products much more quickly, efficiently and at a low cost.

Tellspec’s patented machine learning-based cloud analysis framework can also be used to detect food fraud and food contamination, helping both consumers and authorities to trace problems and prevent food-related health concerns.

This innovative tool allows to detect fish fraud (‘FishQC & Fraud’) in real-time and at a low cost. Fish fraud can occur in the supply chain in different forms: by being adulterated by injecting it with water before it’s frozen; or a species can be substituted and intentionally mislabelled; or farmed fish that is sold as wild fish. Fish fraud is the second biggest fraud in the world after olive oil. The food sensor system allows to measure the actual species of the fish, whether the fish is wild or farmed, or whether the fish is fresh or defrosted in addition to the fish quality (fat, protein, moisture) and spoilage (days of decay), all in a non-destructive process [3].

Tellspec have also developed a related product, Tellspecopedia, which is a search tool that consumers can use to gather objective, evidence-based information on thousands of globally available food ingredients, along with their impact on human health. The database contains all the most common food ingredients, their alternative and chemical names and the health implications of each so you can search up the ingredients you find on labels or typical food components and make informed decisions about the food you buy.

Would you use a food pocket scanner that delivers instantly on your smartphone a breakdown of nutrients before you eat something? Would you use it to make healthier choices about what you eat?