Details of the project EPI

Implementation of innovative methods of managing microorganism cultures at every stage of mead production

01-07-2022 do 31-10-2024

3 585 905,17

EFRROW; national public funds; own funds

https://nvt.pl/arimr/

The aim of the operation is to strengthen the competitiveness of artisan meaderies by implementing innovative technologies, methods of organization and production creating a unique Microorganism Detection System (SDM), ensuring constant and effective monitoring of the microbiological safety of processes at every stage of mead production. Monitoring and evaluation will be carried out in an innovative way, through the use of IT technology using artificial intelligence algorithms to interpret microscopic images remotely, which is a breakthrough on an international scale.

Based on scientific data, the research team will develop a prototype of a microorganism detection system that will automatically answer the question whether the test sample contains microorganisms such as bacteria, yeasts, molds and will determine their numbers. The following actions will be performed: 1. Conducting research in order to develop a methodology for performing microbiological analysis and photos 2. Development of a database of microorganisms along with microscopic photographs 3. Research and development of ML algorithms and models for the analysis of microbiological images 4. Developing an IT system and a mobile application for mead makers

European Commission Regulation (EC) No 2073/2005 of 15 November 2005 on microbiological criteria for foodstuffs requires food producers to carry out microbiological controls during the production process. Standard microbiological inoculations are time-consuming and costly, in addition to often having a high error rate due to underestimation. An additional problem is the transport of samples to the laboratory, which is many times several tens of kilometres away from the mead plant. This is also compounded by the long culture time of the cultures, which results in long waiting times for the results (5-9 days), making it impossible to intervene in the food production process on an ongoing basis. Alternative methods have been sought for many years and include PCR and its variants, Fluorescence in situ Hybridisation, cytometry and many others. There are also intense worldwide attempts to use artificial intelligence as an alternative to microbiological culture by analysing microscopic image

A facilitating aspect is the scientific publications on the application of artificial neural networks in microbial image analysis (Artifcial Intelligence Review https://doi.org/10.1007/s10462-022-10192-7 and Annals of Agricultural and Environmental Medicine 2021, Vol 28, No 4, 705-708) and the possibility of using the results of these studies in the implementation of operations. The main problem to be solved is the detection of microbial cells, their classification into well-defined groups and the counting of the classified microorganisms. Constructing an efficient repository automaton, in other words a deep network architecture, and finding an efficient method for separating and staining micro-organisms from samples, which will allow the resulting material to be used for learning smart networks. A technological challenge is the use of artificial intelligence in the processing of large amounts of data. Such a high-level operation will be the recognition of complex patterns (images) on

SDM will be a fast, low-cost and efficient method for assessing the microbiological quality (safety) of products derived from honey. It will increase the efficiency of the inspection process in companies by providing automated information on the presence of yeasts, bacteria and fungi in a sample in a short time, together with their counts, without the need to deliver samples to a laboratory. This will significantly reduce the cost of controlling production cleanliness by increasing the frequency of routine microbiological inspections and reducing production losses due to exceeding safety standards. In addition, by significantly reducing the cost of testing and testing time, it will be possible to significantly increase testing frequencies, since for the price of one test by traditional methods, 5-10 times as many tests can be performed using SDM. This will have a significant impact on increasing the quality of production. In addition, the SDM system would be available 24/7 at the entrepreneur's disposal, which allows time savings in transporting samples to the laboratory. The system will be easy to use. Each user will have a simple administration panel and a mobile application at their disposal. After a short training course in sampling and sample preparation, each user will be able to take, prepare and play samples into the system using a 'home laboratory'.

The proposed SDM solution will allow fast, automatic, reproducible and low-cost microbiological analysis in accordance with the obligation imposed on food producers by Commission Regulation (EC) No 2073/2005 of 15 November 2005 on microbiological criteria for foodstuffs. SDM is a unique tool on an international scale, based on innovative algorithms using machine-learning mechanisms for the autonomous and automated analysis of microscopic images carried out in a home laboratory. SDM will enable the detection, identification and counting of micro-organisms (bacteria, fungi and yeast) in the food sample under examination. The main advantages of the proposed solution are that it will speed up microbiological analysis time from several days (performing culture tests) to a maximum of two hours. This will increase the efficiency of food production, with producers obtaining test results faster for semi-finished or substrate production, but also for finished products. Producers will not have to hold back the manufactured goods until the results are available, as is the case with culture methods. Such an enormous simplification and shortening of the microbiological purity testing process will also contribute to an increase in the number of tests performed due to increasing the quality of their products. A solution designed to detect and count specific micro-organisms (mesophilic bacteria, yeasts, moulds) has not yet been implemented on the market.

agricultural production system