Electronic nose and computer vision helped roast chicken perfectly

Skoltech research will help restaurants control and automate food preparation processes

in their kitchens.In the future, such technologies may become part of home “smart” ovens. An article detailing the results of this research, supported by a grant from the Russian Science Foundation, was published in the journalFood chemistry.

How do you know if, for example, a chicken is ready to serve?At home, you can watch, smell the food to make sure it is perfectly cooked. However, if you are a chef in a restaurant or a huge industrial kitchen, relying on your eyes and nose to provide consistent, standardized results is difficult. This is why the restaurant industry is actively looking for cheap, reliable and sensitive tools to replace subjective human judgment in order to make quality control automated.

Professor Albert Nasibulin from Skoltech andAalto University, Skoltech Senior Researcher Fyodor Fedorov and their colleagues contributed to solving the problem. They created an "electronic nose" - a set of sensors that detect specific scent components in order to "sniff" the chicken being cooked. In addition, the researchers created a computer vision algorithm so that the system can visually examine the dish. Electronic noses are simpler and cheaper to operate than a gas chromatograph or mass spectrometer, and have even been shown to be capable of detecting different types of cheeses or identifying rotten apples or bananas from a batch of fruit. On the other hand, computer vision is capable of recognizing visual patterns.

Scientists decided to combine these two methods toprecise and non-contact control of the degree of food readiness. They chose chicken meat, popular around the world, and grilled a lot of chicken breasts to "train" the system to evaluate and predict how well it was cooked.

The researchers built their own "electronic nose" with eight sensors and placed it in a ventilation system. The scientists also photographed a grilled chicken and transmitted the information to the algorithm. 

The system was able to identify well enoughundercooked, well-cooked and overcooked chicken, therefore has the potential to automate quality control in the kitchen. The study authors note that in order to use their technique on other parts of the chicken, say, on the legs or wings, or for another cooking method, the electronic "nose" and "eyes" will have to be retrained with new data.

Read more

Abortion and science: what will happen to the children who will give birth

Waste from carrots and beets is made from concrete structures to power generators

Research: humans will not be able to control superintelligent AI machines

The gas chromatograph is a devicefor the analysis of complex gaseous substances by differentiating them into monocomponents. Further, the components of the mixture are analyzed for qualitative and quantitative characteristics. In this case, research can be carried out using any physical and chemical methods.

Mass spectrometry is one of the most accurate methods for identifying substances. In fact, this is a kind of "weighing" of molecules: the components are ionized, then the ratio of the mass to the charge of the ions is determined.