Mammalian nose mimicking AI powered “electronic-nose” to help consumers determine meat freshness. The artificial nose will help identify whether a meat is fit for consumption more accurately than a “Best Before” label could. This technology developed by scientists from Nanyang Technological University, Singapore (NTU Singapore) will reduce wastage of food which is still fit for consumption thereby save the environment.
The electronic-nose has two components:
- Coloured barcode
- AI based barcode reader in the form of a smart phone app.
The barcode will be attached to the inner side of the meat package. As the meat decays it releases certain gases. Upon reacting with the gases released by the meat the barcode will change colour. Using AI the barcode reader will interpret the combinations of colours on the barcode to analyze the freshness of the meat. As claimed by the team within 30 seconds the e-nose can predict the freshness of the meat. In order to make the e-nose system portable the team has integrated the barcode reader in the form of a smart phone app. The barcodes are biodegradable and non-toxic hence can be safely used in food packages.
Scientists have been constantly trying to develop artificial scent screening system also known as electronic nose. However few e-noses have made to the commercial level. The e-noses developed in the past have failed mainly due to sensing and scent fingerprint recognition problems. They could not precisely identify and determine the odour which was a great disadvantage. These problems have arisen in the previous e-nose approaches because of lack of robust cross-reactive sensing and scent fingerprint recognition. The NTU team has developed a quick, automatic, accurate, portable, real-time artificial scent screening system or electronic nose (e-nose). NTU team’s e-nose has overcome the problems of its predecessors. Their e-nose could be used to sniff the freshness of packaged food items just like we human beings sniff food to check its fitness for consumption. The team’s ground breaking work has been published in the journal Advanced Materials.
The NTU team developed e-nose functions similar to a mammalian nose:
- Functioning of the mammalian nose: In case of a mammalian nose gases produced by rotting meat bind to the receptors in the nose. This generates signals which get transmitted to the brain. After receiving the signal the brain arranges the response in the form of a pattern. This allows the mammal to identify the odour and understand whether the meat is rotting or fresh.
- Functioning of the e-nose: Imitating the mammalian nose mechanism the e-nose has been developed by the team. In case of the e-nose the barcode has 20 bars which function as the nose receptor. The chitosan (natural sugar) made bars, each loaded with a specific dye, are embedded on a cellulose derivative. The gases produced by the decaying meat react with the dye in the bar which changes the colour of the dye. Depending on the concentration and type of gases emitted by the meat a unique combination of colour is produced in the barcode. This colour combination acts as the “scent fingerprint” for the condition of the meat. Then the AI based barcode indicator will interpret the unique colour combination to predict the freshness of the meat.
The team has demonstrated how the barcode component of the e-nose works. The first bar, for example, contains a weakly acidic yellow dye. As the meat decays it produces nitrogen containing compounds called biomanines. Upon exposure to the bioamines the yellow dye reacts with the compounds. The colour of the dye changes from yellow to blue due to the reaction. Overtime as the meat rots further there is an increase in production of bioamines. The higher concentration of bioamines further intensifies the colour. The rest of the bars in the barcode function similarly. Using an international standard the research team developed a classification system to define meat freshness. The three classifications for meat freshness was: fresh, less fresh and spoiled.
Plenty of food items which are fit for consumption are discarded unknowingly. People can save money by using this mammalian nose mimicking e-nose to determine whether the meat is fit for consumption irrespective of what is mentioned on Best Before labels.
The NTU team tested the effectiveness of their e-nose on commercially available beef, chicken and fish packages. In one such experimental study the team continuously monitored the freshness of fish packed with commonly used transparent PVC films over five days. The study fish was stored at 4°C during the five day time period. The team monitored the freshness of the fish, without touching the fish, using their ingenious barcode attached to the inner surface of the PVC film. The barcode measured the ammonia and two other bioamines released by the fish over time. The team clicked images of the barcode at different intervals over the study period. The barcode reader developed by the research team is based on deep convolutional neural networks, a type of AI algorithm. With the help of the images of different barcodes the team trained the AI powered barcode to interpret the “scent fingerprint” patterns which would be used to analyze and predict the freshness of the meat. The team observed their novel e-nose was capable of monitoring the freshness of the meat with 98.5% accuracy.
The team’s proof of concept e-nose system has been tested successfully in real life scenarios. Their e-nose has overcome the hurdle of previously developed e-nose systems which used bulky wiring system to collect electrical signals. According to the team the portable artificial olfactory system which yields results in a short time can easily be integrated in food packaging materials. Being biodegradable and non toxic the novel barcode can safely be used in all parts of the food supply chain system to monitor freshness of packaged food items.
The team has filed a patent for this portable artificial olfactory system for measuring food freshness in real-time. In order to extend this novel technology to other types of perishables the team is working with a Singapore agribusiness company. By further developing this e-nose to recognise the types of gases which are produced by other perishable food items as they start to decay this technology will have a wider application in food quality control.
This non-destructive, fast, automated, accurate and real-time robust food freshness monitoring technology will help consumers save money and the environment by reducing food wastage. The AI powered e-nose will enable consumers and stake holders in the food supply chain to monitor food freshness.
- Lingling Guo, Ting Wang, Zhonghua Wu, Jianwu Wang, Ming Wang, Zequn Cui, Shaobo Ji, Jianfei Cai, Chuanlai Xu, Xiaodong Chen. Portable Food‐Freshness Prediction Platform Based on Colorimetric Barcode Combinatorics and Deep Convolutional Neural Networks. Advanced Materials, 2020; 32 (45): 2004805 DOI: 10.1002/adma.202004805
Available from: https://onlinelibrary.wiley.com/doi/10.1002/adma.202004805