MIL OSI Translation. Region: Russian Federation –
Source: Novosibirsk State University – Novosibirsk State University –
Scientists from NSU and ICG SB RAS presented a new approach to collecting, storing and analyzing information on the morphometric characteristics of a wheat ear. Students took an active part in the work on creating the SpikeDroidDB system Faculty of Mechanics and Mathematics of NSU, Faculty of Information Technology NSU, and also Mathematical center in AkademgorodokWork on this project was carried out with the support of the Russian Science Foundation, project No. 23-14-00150.
The SpikeDroidDB information system allows storing digital images of the ear, annotating their phenotypic characteristics according to 14 important traits and provides a flexible query system for accessing data.
Using SpikeDroidDB, a collection of F2 hybrid ears from a cross between the Australian soft wheat variety Triple Dirk and the Chinese wheat sample KU506 Triticum yunnanense was digitized and annotated. An analysis of the variability of the ears in shape, length and width was carried out.
The structure of the ear is one of the most important features of cereals, associated with such economically valuable qualities as productivity, resistance to environmental factors and pests, ease of threshing. Ears differ in shape, size, density, awns, color, etc.
For breeders and geneticists, such parameters as the number of grains in an ear, the thousand-grain weight, and others are of great importance. These characteristics are closely related to plant productivity. A useful selection feature is the shape of the grain and such characteristics of the ear as its type, length, profile, the presence or absence of awns, the number of fertile and sterile spikelets (i.e., grain content), ear fragility, and the properties of the glume. Collecting and describing these features manually is a labor-intensive and lengthy process.
— Researchers at our laboratory have long been working on an important task aimed at replacing the measuring methods of geneticists and breeders with a ruler and a computer or mobile phone. We would like to make it so that scientists no longer have to manually measure plant parameters, but simply take a photo of a wheat ear, while observing a number of technical conditions, and then obtain the information they are interested in by uploading this photo to our database. When creating it, we worked with conventional image analysis, that is, with digital vision, and applied deep machine learning in terms of image recognition using neural networks, identifying individual features and classification, — said the leading researcher at the Laboratory of Evolutionary Informatics and Theoretical Genetics of the Institute of Cytology and Genetics of the Siberian Branch of the Russian Academy of Sciences, an employee of the Department of Information Biology Faculty of Natural Sciences, Novosibirsk State University Dmitry Afonnikov.
The complexity of the researchers’ work was that it was based on digital images of ears. They serve as the initial data when applying automatic phenotyping methods. When developing them, an important task is an expert assessment of many plant characteristics for their further use in training and verification of computer algorithms. However, many morphological features of the ear are usually assessed qualitatively, not quantitatively. Very often, they do not have a quantitative assessment. Such features include the shape of the ear, its density, the color of the ear, the pubescence of the glumes, the type of awns, the color of the awns, the shape of the ear, brittleness of the ear and many others. Therefore, the use of digital image analysis approaches to describe the shape of the grain and ear, as well as their comparison with the assessments of the ear features made by expert breeders, became an important task for the developers.
— In our database, we have collected over 10,000 digital images of ears and described their structure and properties so that genetic scientists can obtain all the data they need from a photograph — the size of the ear, its thickness, width, presence of awns, color of the ears, etc., essentially replacing conventional measurements with image analysis. And as a result, we obtain more characteristics, and they are also more accurate. In this case, the automated system has more capabilities than a person. If a person determines some parameters “by eye”, then computer vision records them more accurately and productively. With the help of computer analysis of digital images, we can determine hundreds of parameters of ears — both basic and their derivatives, and then use them to develop methods and classifications, as well as to assess productivity. Such technologies provide a high degree of automation of information collection, its storage in databases, integration with data on the genotype and environmental parameters, and create the basis for intelligent analysis of the information received. There is another important advantage: a digital description of the ear and its image will be stored in the database for as long as necessary, whereas a dried ear placed in a paper envelope may crumble, change color or deteriorate, and the sample will be lost, explained Dmitry Afonnikov.
In the SpikeDroidDB system, several images can be associated with each ear. For each of them, the protocol by which it was obtained is indicated. For shooting, the developers used two protocols for obtaining digital images of mature ears. They chose a blue background as the most contrasting to the color of the ears and allowing you to easily separate the object from the background. Shooting of the ears was carried out in two versions: in the first, the ear is located vertically in front of a blue background, the second shooting option provides for a horizontal position of the ears on the glass above the blue background.
The prototype of the SpikeDroidDB system is available at this link http://speakedroid.biores.cytogen.ru/The main page contains brief information about the database, links for logging in or registering, and links to the main blocks of information in the database.
Dmitry Afonnikov says that breeders and geneticists involved in developing new varieties of wheat are showing great interest in this development and are interested in working with it to automate painstaking and lengthy routine processes that require precision and concentration. In addition, the SpikeDroidDB system will help avoid subjective assessments, errors and inaccuracies in phenotyping ear samples.
Please note: This information is raw content directly from the source of the information. It is exactly what the source states and does not reflect the position of MIL-OSI or its clients.
Please note; This information is raw content directly from the information source. It is accurate to what the source is stating and does not reflect the position of MIL-OSI or its clients.
https://www.nsu.ru/n/media/news/nauka/uchenye-ngu-i-itsig-so-ran-razrabotali-novyy-podkhod-dlya-sbora-khraneniya-i-analiza-informatsii-ok/
EDITOR’S NOTE: This article is a translation. Apologies should the grammar and or sentence structure not be perfect.