Food Processing: Techniques and Technology Food Processing: Techniques and Technology Техника и технология пищевых производств 2074-9414 2313-1748 69869 10.21603/2074-9414-2023-3-2448 ОРИГИНАЛЬНАЯ СТАТЬЯ ORIGINAL ARTICLE ОРИГИНАЛЬНАЯ СТАТЬЯ Automated Measurement of Air Bubbles Dispersion in Ice Cream Using Machine Learning Methods Автоматизированное определение дисперсности воздушной фазы в мороженом с применением методов машинного обучения https://orcid.org/0000-0003-3166-2827 Королев Игорь Антонович Korolev Igor A. i.korolev@fncps.ru Всероссийский научно-исследовательский институт холодильной промышленности Москва Россия All-Russian Scientific Research Institute of Refrigeration Industry Moscow Russian Federation 29 09 2023 29 09 2023 53 3 455 464 29 11 2022 10 01 2023 https://fptt.ru/en/issues/22025/21971/

Мороженое – популярный замороженный десерт, характеризующийся присутствием воздушной фазы в виде мельчайших пузырьков (средний диаметр 15–60 мкм). При создании новых рецептур мороженого необходимо учитывать влияние композиционного состава и производственных факторов на состояние воздушной фазы. Важно достоверно и с минимальными затратами времени определять ее дисперсность. Цель исследования – разработка и оптимизация алгоритма для автоматической разметки положения центров и диаметра воздушных пузырьков на микроскопических изображениях типа ограничивающая окружность на примере мороженого, а также создание компьютерной программы. Провел поиск и анализ статей по применению микроскопического метода исследований для оптимизации технологических параметров работы оборудования или рецептуры мороженого и замороженных десертов, опубликованных на русском и английском языках в базах данных WoS и RSCI за последние 20 лет. Выполнил сбор микроскопических изображений воздушной фазы в мороженом при помощи микроскопа Оlympus CX41RF. Для написания автоматической программы разметки использовал язык программирования Python, библиотеку машинного обучения Keras и фреймворк TensorFlow. Обучение моделей выполнено при помощи видеоускорителя NVIDIA GTX. Анализ научной литературы показал, что дисперсность воздушной фазы в мороженом зависит от его состава и параметров процесса фризерования и закаливания, а формирование воздушных пузырьков описывается теориями пенообразования. На основе ручной разметки микроскопических изображений пузырьков воздуха в мороженом создал обучающий массив данных. Определил оптимальное количество каналов в конволюционных слоях нейронной сети на базе архитектуры LeNet, которая позволила классифицировать изображения сфер/не сфер с точностью более 0,995. Установил пределы срабатывания нейронной сети при использовании метода скользящего окна: 7,5 % диаметра при боковом смещении, 12,5 % при масштабировании. Разработанный алгоритм позволяет автоматически размечать пузырьки на микроскопических изображениях. Погрешность определения среднего диаметра сфер не превышала 1,8 %. Разработал метод автоматизированного подсчета количества и диаметра пузырьков воздуха в мороженом, который позволяет снизить трудоемкость проводимых исследований. Исходный код проекта размещен в открытом доступе и может быть адаптирован исследователями для решения своих задач в области компьютерного зрения.

Ice cream is a popular cold dessert. Its air phase consists of tiny bubbles with an average diameter of 15–60 µm. New ice cream formulations depend on the way the composition and production factors affect the air phase. As a result, ice cream producers need new time-saving and reliable methods to determine dispersion. The research objective was to create a computer program for marking the position of centers and diameter of air bubbles on microscopic images of a bounding circle type. The review part included 20 years of Russian and English publications on microscopic research methods in ice cream production indexed in Web of Science and Russian Research Citation Index. Microscopic images of ice cream air phase were obtained using an Olympus CX41RF microscope with a magnification of ×100. The automatic markup program employed the Python programming language, the Keras machine learning library, and the TensorFlow framework. The models were trained using the NVIDIA GTX video accelerator. The review showed that the dispersion of ice cream air phase depends on its composition and the freezing parameters whereas bubble formation is usually described in line with the existing foaming theories. A training data set was obtained by manual labeling of microscopic images. The optimal number channels in the convolutional layers of a neural network with LeNet-type architecture was determined, which made it possible to classify images as spheres or non-spheres with an accuracy of ≥ 0.995. The sliding window method helped to determine the limits of the neural network triggering for the sliding window method were determined, which reached 7.5% of the diameter with lateral displacement and 12.5% with scaling. The developed algorithm automatically marked bubbles on microscopic images. The error in determining the average diameter was below 1.8%. The new method for automated calculation of the number and diameter of air bubbles in ice cream proved to be user-friendly. It can be found in public domain, and researchers are free to adapt it to solve various computer vision issues.

 Мороженое дисперсность пузырьки воздуха микроскопия машинное обучение нейронная сеть Ice cream microscopy machine learning markup bounding circle air bubbles Статья подготовлена в рамках выполнения исследований по государственному заданию Федерального научного центра пищевых систем им. В. М. Горбатова РАН. The article was part of state assignment to the V.M. Gorbatov Federal Research Center for Food Systems of Russian Academy of Science.

Genovese A, Balivo A, Salvati A, Sacchi R. Functional ice cream health benefits and sensory implications. Food Research International. 2022;161. https://doi.org/10.1016/j.foodres.2022.111858 Genovese A, Balivo A, Salvati A, Sacchi R. Functional ice cream health benefits and sensory implications. Food Research International. 2022;161. https://doi.org/10.1016/j.foodres.2022.111858 Akbari M, Eskandari MH, Davoudi Z. Application and functions of fat replacers in low-fat ice cream: A review. Trends in Food Science and Technology. 2019;86:34-40. https://doi.org/10.1016/j.tifs.2019.02.036 Akbari M, Eskandari MH, Davoudi Z. Application and functions of fat replacers in low-fat ice cream: A review. Trends in Food Science and Technology. 2019;86:34-40. https://doi.org/10.1016/j.tifs.2019.02.036 Gheisari HR, Heydari S, Basiri S. The effect of date versus sugar on sensory, physicochemical, and antioxidant properties of ice cream. Iranian Journal of Veterinary Research. 2020;21(1):9-14. Gheisari HR, Heydari S, Basiri S. The effect of date versus sugar on sensory, physicochemical, and antioxidant properties of ice cream. Iranian Journal of Veterinary Research. 2020;21(1):9-14. Arslaner A, Salik MA. Functional ice cream technology. Akademik Gıda. 2020;18(2):180-189. https://doi.org/10.24323/akademik-gida.758835 Arslaner A, Salik MA. Functional ice cream technology. Akademik Gıda. 2020;18(2):180-189. https://doi.org/10.24323/akademik-gida.758835 Ademosun AO. Glycemic properties of soursop-based ice cream enriched with moringa leaf powder. Foods and Raw Materials. 2021;9(2):207-214. https://doi.org/10.21603/2308-4057-2021-2-207-214 Ademosun AO. Glycemic properties of soursop-based ice cream enriched with moringa leaf powder. Foods and Raw Materials. 2021;9(2):207-214. https://doi.org/10.21603/2308-4057-2021-2-207-214 Wu B, Freire DO, Hartel RW. The effect of overrun, fat destabilization, and ice cream mix viscosity on entire meltdown behavior. Journal of Food Science. 2019;84(9):2562-2571. https://doi.org/10.1111/1750-3841.14743 Wu B, Freire DO, Hartel RW. The effect of overrun, fat destabilization, and ice cream mix viscosity on entire meltdown behavior. Journal of Food Science. 2019;84(9):2562-2571. https://doi.org/10.1111/1750-3841.14743 Cheng J, Ma Y, Li X, Yan T, Cui J. Effects of milk protein-polysaccharide interactions on the stability of ice cream mix model systems. Food Hydrocolloids. 2015;45:327-336. https://doi.org/10.1016/j.foodhyd.2014.11.027 Cheng J, Ma Y, Li X, Yan T, Cui J. Effects of milk protein-polysaccharide interactions on the stability of ice cream mix model systems. Food Hydrocolloids. 2015;45:327-336. https://doi.org/10.1016/j.foodhyd.2014.11.027 da Silva Faresin L, Devos RJB, Reinehr CO, Colla LM. Development of ice cream with reduction of sugar and fat by the addition of inulin, Spirulina platensis or phycocyanin. International Journal of Gastronomy and Food Science. 2022;27. https://doi.org/10.1016/j.ijgfs.2021.100445 da Silva Faresin L, Devos RJB, Reinehr CO, Colla LM. Development of ice cream with reduction of sugar and fat by the addition of inulin, Spirulina platensis or phycocyanin. International Journal of Gastronomy and Food Science. 2022;27. https://doi.org/10.1016/j.ijgfs.2021.100445 Chang Y, Hartel RW. Development of air cells in a batch ice cream freezer. Journal of Food Engineering. 2002;55(1):71-78. https://doi.org/10.1016/S0260-8774(01)00243-6 Chang Y, Hartel RW. Development of air cells in a batch ice cream freezer. Journal of Food Engineering. 2002;55(1):71-78. https://doi.org/10.1016/S0260-8774(01)00243-6 Warren MM, Richard WH. Structural, compositional, and sensorial properties of United States commercial ice cream products. Journal of Food Science. 2014;79(10):E2005-E2013. https://doi.org/10.1111/1750-3841.12592 Warren MM, Richard WH. Structural, compositional, and sensorial properties of United States commercial ice cream products. Journal of Food Science. 2014;79(10):E2005-E2013. https://doi.org/10.1111/1750-3841.12592 Yan G, Cui Y, Lia D, Ding Y, Han J, Wang S, et al. The characteristics of soybean protein isolate obtained by synergistic modification of high hydrostatic pressure and phospholipids as a promising replacement of milk in ice cream. LWT. 2022;160. https://doi.org/10.1016/j.lwt.2022.113223 Yan G, Cui Y, Lia D, Ding Y, Han J, Wang S, et al. The characteristics of soybean protein isolate obtained by synergistic modification of high hydrostatic pressure and phospholipids as a promising replacement of milk in ice cream. LWT. 2022;160. https://doi.org/10.1016/j.lwt.2022.113223 Исследование воздушной фазы мороженого, содержащего НФ-концентрат творожной сыворотки / В. Н. Шохалова [и др.] // Молочнохозяйственный вестник. 2017. Т. 26. № 2. С. 130-137. https://www.elibrary.ru/ZBFZXF Shokhalova VN, Kusin AA, Shokhalov VA, Kostyukov EM. Investigation of ice-cream air phase with a curd whey NF concentrate. Molochnokhozayistvenny Vestnik. 2017;26(2):130-137. (In Russ.). https://www.elibrary.ru/ZBFZXF Gurskiy IA, Tvorogova AA, Shobanova TV. The condition of the structure of the thawed aerated sour-milk desserts during its storage. Proceedings of the Voronezh State University of Engineering Technologies. 2020;82(2):94-100. (In Russ.). https://doi.org/10.20914/2310-1202-2020-2-94-100 Gurskiy IA, Tvorogova AA, Shobanova TV. The condition of the structure of the thawed aerated sour-milk desserts during its storage. Proceedings of the Voronezh State University of Engineering Technologies. 2020;82(2):94-100. (In Russ.). https://doi.org/10.20914/2310-1202-2020-2-94-100 Landikhovskaya AV, Tvorogova AA, Kazakova NV, Gursky IA. The effect of trehalose on dispersion of ice crystals and consistency of low-fat ice cream. Food Processing: Techniques and Technology. 2020;50(3):450-459. (In Russ.). https://doi.org/10.21603/2074-9414-2020-3-450-459 Landikhovskaya AV, Tvorogova AA, Kazakova NV, Gursky IA. The effect of trehalose on dispersion of ice crystals and consistency of low-fat ice cream. Food Processing: Techniques and Technology. 2020;50(3):450-459. (In Russ.). https://doi.org/10.21603/2074-9414-2020-3-450-459 Gurskiy IA, Tvorogova AA. The effect of whey protein concentrates on technological and sensory quality indicators of ice cream. Food Processing: Techniques and Technology. 2022;52(3):439-448. (In Russ.). https://doi.org/10.21603/2074-9414-2022-3-2376 Gurskiy IA, Tvorogova AA. The effect of whey protein concentrates on technological and sensory quality indicators of ice cream. Food Processing: Techniques and Technology. 2022;52(3):439-448. (In Russ.). https://doi.org/10.21603/2074-9414-2022-3-2376 Jing X, Chen Z, Tang Z, Tao Y, Huang Q, Wu Y, et al. Preparation of camellia oil oleogel and its application in an ice cream system. LWT. 2022;169. https://doi.org/10.1016/j.lwt.2022.113985 Jing X, Chen Z, Tang Z, Tao Y, Huang Q, Wu Y, et al. Preparation of camellia oil oleogel and its application in an ice cream system. LWT. 2022;169. https://doi.org/10.1016/j.lwt.2022.113985 Schroeder AB, Dobson ETA, Rueden CT, Tomancak P, Jug F, Eliceiri KW. The ImageJ ecosystem: Open-source software for image visualization, processing, and analysis. Protein Science. 2021;30(1):234-249. https://doi.org/10.1002/pro.3993 Schroeder AB, Dobson ETA, Rueden CT, Tomancak P, Jug F, Eliceiri KW. The ImageJ ecosystem: Open-source software for image visualization, processing, and analysis. Protein Science. 2021;30(1):234-249. https://doi.org/10.1002/pro.3993 Ilonen J, Juránek R, Eerola T, Lensu L, Dubská M, Zemčík P, et al. Comparison of bubble detectors and size distribution estimators. Pattern Recognition Letters. 2018;101:60-66. https://doi.org/10.1016/j.patrec.2017.11.014 Ilonen J, Juránek R, Eerola T, Lensu L, Dubská M, Zemčík P, et al. Comparison of bubble detectors and size distribution estimators. Pattern Recognition Letters. 2018;101:60-66. https://doi.org/10.1016/j.patrec.2017.11.014 Cui Y, Li C, Zhang W, Ning X, Shi X, Gao J, et al. A deep learning-based image processing method for bubble detection, segmentation, and shape reconstruction in high gas holdup sub-millimeter bubbly flows. Chemical Engineering Journal. 2022;449. https://doi.org/10.1016/j.cej.2022.137859 Cui Y, Li C, Zhang W, Ning X, Shi X, Gao J, et al. A deep learning-based image processing method for bubble detection, segmentation, and shape reconstruction in high gas holdup sub-millimeter bubbly flows. Chemical Engineering Journal. 2022;449. https://doi.org/10.1016/j.cej.2022.137859 Cerqueira RFL, Paladino EE. Development of a deep learning-based image processing technique for bubble pattern recognition and shape reconstruction in dense bubbly flows. Chemical Engineering Science. 2021;230. https://doi.org/10.1016/j.ces.2020.116163 Cerqueira RFL, Paladino EE. Development of a deep learning-based image processing technique for bubble pattern recognition and shape reconstruction in dense bubbly flows. Chemical Engineering Science. 2021;230. https://doi.org/10.1016/j.ces.2020.116163 Jabarkhyl S, Barigou M, Zhu S, Rayment P, Lloyd DM, Rossetti D. Foams generated from viscous non-Newtonian shear-thinning liquids in a continuous multi rotor-stator device. Innovative Food Science and Emerging Technologies. 2020;59. https://doi.org/10.1016/j.ifset.2019.102231 Jabarkhyl S, Barigou M, Zhu S, Rayment P, Lloyd DM, Rossetti D. Foams generated from viscous non-Newtonian shear-thinning liquids in a continuous multi rotor-stator device. Innovative Food Science and Emerging Technologies. 2020;59. https://doi.org/10.1016/j.ifset.2019.102231 Cook KLK, Hartel RW. Mechanisms of ice crystallisation in ice cream production. Comprehensive Reviews in Food Science and Food Safety. 2010;9(2):213-222. https://doi.org/10.1111/j.1541-4337.2009.00101.x Cook KLK, Hartel RW. Mechanisms of ice crystallisation in ice cream production. Comprehensive Reviews in Food Science and Food Safety. 2010;9(2):213-222. https://doi.org/10.1111/j.1541-4337.2009.00101.x Caillet A, Cogné C, Andrieu J, Laurent P, Rivoire A. Characterization of ice cream structure by direct optical microscopy. Influence of freezing parameters. LWT - Food Science and Technology. 2003;36(8):743-749. https://doi.org/10.1016/S0023-6438(03)00094-X Caillet A, Cogné C, Andrieu J, Laurent P, Rivoire A. Characterization of ice cream structure by direct optical microscopy. Influence of freezing parameters. LWT - Food Science and Technology. 2003;36(8):743-749. https://doi.org/10.1016/S0023-6438(03)00094-X Eisner MD, Wildmoser H, Windhab EJ. Air cell microstructuring in a high viscous ice cream matrix. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2005;263(1-3):390-399. https://doi.org/10.1016/j.colsurfa.2004.12.017 Eisner MD, Wildmoser H, Windhab EJ. Air cell microstructuring in a high viscous ice cream matrix. Colloids and Surfaces A: Physicochemical and Engineering Aspects. 2005;263(1-3):390-399. https://doi.org/10.1016/j.colsurfa.2004.12.017 Parra ODH, Ndoye F-T, Benkhelifa H, Flick D, Alvarez G. Effect of process parameters on ice crystals and air bubbles size distributions of sorbets in a scraped surface heat exchanger. International Journal of Refrigeration. 2018;92:225-234. https://doi.org/10.1016/j.ijrefrig.2018.02.013 Parra ODH, Ndoye F-T, Benkhelifa H, Flick D, Alvarez G. Effect of process parameters on ice crystals and air bubbles size distributions of sorbets in a scraped surface heat exchanger. International Journal of Refrigeration. 2018;92:225-234. https://doi.org/10.1016/j.ijrefrig.2018.02.013 Goraya RK, Singla M, Bajwa U, Kaur A, Pathania S. Impact of sodium alginate gelling and ingredient amalgamating order on ingredient interactions and structural stability of ice cream. LWT. 2021;147. https://doi.org/10.1016/j.lwt.2021.111558 Goraya RK, Singla M, Bajwa U, Kaur A, Pathania S. Impact of sodium alginate gelling and ingredient amalgamating order on ingredient interactions and structural stability of ice cream. LWT. 2021;147. https://doi.org/10.1016/j.lwt.2021.111558 Akalın AS, Kesenkas H, Dinkci N, Unal G, Ozer E, Kınık O. Enrichment of probiotic ice cream with different dietary fibers: Structural characteristics and culture viability. Journal of Dairy Science. 2018;101(1):37-46. https://doi.org/10.3168/jds.2017-13468 Akalın AS, Kesenkas H, Dinkci N, Unal G, Ozer E, Kınık O. Enrichment of probiotic ice cream with different dietary fibers: Structural characteristics and culture viability. Journal of Dairy Science. 2018;101(1):37-46. https://doi.org/10.3168/jds.2017-13468 Samakradhamrongthai RS, Jannu T, Supawan T, Khawsud A, Aumpa P, Renaldi G. Inulin application on the optimization of reduced-fat ice cream using response surface methodology. Food Hydrocolloids. 2021;119. https://doi.org/10.1016/j.foodhyd.2021.106873 Samakradhamrongthai RS, Jannu T, Supawan T, Khawsud A, Aumpa P, Renaldi G. Inulin application on the optimization of reduced-fat ice cream using response surface methodology. Food Hydrocolloids. 2021;119. https://doi.org/10.1016/j.foodhyd.2021.106873 Narala VR, Orlovs I, Jugbarde MA, Masin M. Inulin as a fat replacer in pea protein vegan ice cream and its influence on textural properties and sensory attributes. Applied Food Research. 2022;2(1). https://doi.org/10.1016/j.afres.2022.100066 Narala VR, Orlovs I, Jugbarde MA, Masin M. Inulin as a fat replacer in pea protein vegan ice cream and its influence on textural properties and sensory attributes. Applied Food Research. 2022;2(1). https://doi.org/10.1016/j.afres.2022.100066 Baldominos A, Saez Y, Isasi P. A survey of handwritten character recognition with MNIST and EMNIST. Applied Sciences. 2019;9(15). https://doi.org/10.3390/app9153169 Baldominos A, Saez Y, Isasi P. A survey of handwritten character recognition with MNIST and EMNIST. Applied Sciences. 2019;9(15). https://doi.org/10.3390/app9153169 Rosebrock A. Deep learning for computer vision with python: Starter bundle. PyImageSearch; 2017. 330 p. Rosebrock A. Deep learning for computer vision with python: Starter bundle. PyImageSearch; 2017. 330 p. Prashanth DS, Mehta RVK, Ramana K, Bhaskar V. Handwritten Devanagari Character Recognition using modified lenet and alexnet convolution neural networks. Wireless Personal Communications. 2022;122:349-378. https://doi.org/10.1007/s11277-021-08903-4 Prashanth DS, Mehta RVK, Ramana K, Bhaskar V. Handwritten Devanagari Character Recognition using modified lenet and alexnet convolution neural networks. Wireless Personal Communications. 2022;122:349-378. https://doi.org/10.1007/s11277-021-08903-4