The fruit fly Drosophila melanogaster is a classic research object in genetics and systems biology. In the genetic analysis of flies, a routine task is to determine the offspring size and gender ratio in their populations. Currently, these estimates are made manually, which is a very time-consuming process. The counting and gender determination of flies can be automated by using image analysis with deep learning neural networks on mobile devices. We proposed an algorithm based on the YOLOv4-tiny network to identify Drosophila flies and determine their gender based on the protocol of taking pictures of insects on a white sheet of paper with a cell phone camera. Three strategies with different types of augmentation were used to train the network. The best performance (F1 = 0.838) was achieved using synthetic images with mosaic generation. Females gender determination is worse than that one of males. Among the factors that most strongly influencing the accuracy of fly gender recognition, the fly’s position on the paper was the most important. Increased light intensity and higher quality of the device cameras have a positive effect on the recognition accuracy. We implement our method in the FlyCounter Android app for mobile devices, which performs all the image processing steps using the device processors only. The time that the YOLOv4-tiny algorithm takes to process one image is less than 4 s.

No abstract

The structure of the ear is one of the most important features of cereals associated with such agronomically important traits as productivity, resistance to environ mental factors and pests, threshebility. Ears differ in shape, size, density, awnedness, color, etc. Analysis of the ear traits requires visual inspection, manual measurements and is very timeconsuming. The ef fec tiveness of ears' phenotyping can be improved by the introduction of an automated image process ing technology, storage of information in databases, use of machine learning algorithms to analyze this information. This paper presents a new approach for collecting, storing and analyzing of information about morphometric characteristics of ears of wheat. Two protocols for obtaining digital images of the ear have been developed. The computeraided information system SpikeDroidDB has been developed, which allows you to store digital images of the ear, annotate their phenotypic features (14 features, including plant variety description, links to parent genotypes, genera tion, planting number, ear morphology description). The interface provides a flexible query system to access the data. SpikeDroidDB represents an intercon nected representation between genotype, phenotype, location, and growing conditions. The web interface of SpikeDroidDB is available at http://spikedroid.biores. cytogen.ru/ and allows you to work with the system as with desktop computers or mobile devices. We used SpikeDroidDB for the digitization and annotation of a collection of ears of F 2 hybrids from crosses between the Australian cultivar of common wheat Triple Dirk and accession KU506 of Chinese wheat Triticum yunnanense. This experiment includes analysis of 104 plants, 230 spike images. The analysis of the variability of ears in form, length, and other traits allowed determination of the type of their genetic control: compactness is controlled by two recessive genes, awn type and hairi Структура колоса -один из важнейших признаков злаков, связан ный с такими их хозяйственно ценными качествами, как продук тивность, устойчивость к факторам внешней среды и вредителям, легкость обмолота. Колосья различаются по форме, размерам, плотности, остистости, цвету и т. д. Оценка характеристик колоса выполняется экспертом на основании визуального анализа и тре бует существенных затрат времени. Эффективность фенотипиро вания колосьев можно повысить за счет внедрения компьютерных технологий, организации хранения информации в базах данных, использования алгоритмов машинного обучения для анализа по лученной информации. В настоящей работе представлен новый подход для сбора, хранения и анализа информации о морфомет рических характеристиках колоса пшеницы. Разработано несколь ко протоколов получения цифровых изображений колоса. Созда на компьютерная информационная система SpikeDroidDB, которая позволяет хранить цифровые изображения колоса, аннотировать их фенотипические характеристики (всего 14 признаков), предо ставляет гибкую систему запросов для доступа к данным. В систе ме SpikeDroidDB для...

No abstract