Search

Abstract Automatic detection of tomato leaf spot disease is essential for control and loss reduction. Traditional algorithms face challenges such as large amount of data, multiple training and heavy computation. In this study, a lightweight shared Siamese neural network method was proposed for tomato leaf disease identification, which is suitable for resource‐limited environments. Experiments on Plant‐Village, Taiwan and Taiwan ++ datasets show that the accuracy fluctuates very little even when trained with only 60% of the data, which confirms the effectiveness of the proposed method in the small data environment. In addition, compared with the mainstream algorithms, it improves the accuracy by up to 35.3%on Plant‐Village and two Taiwan datasets respectively. The experimental results also show that the proposed method still performs well when the data is imbalanced and the sample size is small.

Abstract This paper introduces few‐shot anomaly detection (FSAD), a practical and less anomaly detection (AD) method, which can provide a limited number of normal images for each class during training. So far, studies on FSAD have been carried out according to each model, and there is no discussion of commonalities between different types. Depending on how people detect unusual lies, the problematic images are compared to the normal ones. The image alignment method based on different classifications is used to train the target detection model independent of classification, and performed ablation experiments on the pest and disease datasets in different environments for verification. This is the first time the FSAD method has been used to train a single scalable model without the need to train new classifications or adjust parameters. The experimental results show that the application of AUC based on vegetation disease data set and vegetation pest data set in FSAD algorithm is improved by 19.5% compared with the existing algorithm.

It is of great significance to identify the pest species accurately and control it effectively to reduce the loss of agricultural products. The research results of this project will provide theoretical basis for preventing and controlling the spread of pests and reducing the loss of agricultural products, and have important practical significance for improving the quality of agricultural products and increasing the output of agricultural products. At the same time, it provides a kind of effective prevention and control measures for farmers, so as to ensure the safety and health of crops. Because of the slow speed and high cost of manual identification, it is necessary to establish a set of automatic pest identification system. The traditional image-based insect classifier is mainly realized by machine vision technology, but because of its high complexity, the classification efficiency is low and it is difficult to meet the needs of applications. Therefore, it is necessary to develop a new automatic insect recognition system to improve the accuracy of insect classification. There are many species and forms of insects, and the field living environment is complex. The morphological similarity between species is high, which brings difficulties to the classification of insects. In recent years, with the rapid development of deep learning technology, using artificial neural network to classify pests is an important method to establish a fast and accurate classification model. In this work, we propose a novel convolutional neural network-based model (MSSN), which includes attention mechanism, feature pyramid, and fine-grained model. The model has good scalability, can better capture the semantic information in the image, and achieve more accurate classification. We evaluated our approach on a common data set: large-scale pest data set, PlantVillage benchmark data set, and evaluated model performance using a variety of evaluation indicators, namely, macro mean accuracy (MPre), macro mean recall rate (MRec), macro mean F1-score (MF1), Accuracy (Acc) and geometric mean (GM). Experimental results show that the proposed algorithm has better performance and universality ability than the existing algorithm. For example, on the data set, the maximum accuracy we obtained was 86.35%, which exceeded the corresponding technical level. The ablation experiment was conducted on the experiment itself, and the comprehensive evaluation of the complete MSSN(scale 1+2+3) was the best in various performance indexes, demonstrating the feasibility of the innovative method in this paper.

BackgroundThe process of crop domestication often consists of two stages: initial domestication, where the wild species is first cultivated by humans, followed by diversification, when the domesticated species are subsequently adapted to more environments and specialized uses. Selective pressure to increase sugar accumulation in certain varieties of the cereal crop Sorghum bicolor is an excellent example of the latter; this has resulted in pronounced phenotypic divergence between sweet and grain-type sorghums, but the genetic mechanisms underlying these differences remain poorly understood.ResultsHere we present a new reference genome based on an archetypal sweet sorghum line and compare it to the current grain sorghum reference, revealing a high rate of nonsynonymous and potential loss of function mutations, but few changes in gene content or overall genome structure. We also use comparative transcriptomics to highlight changes in gene expression correlated with high stalk sugar content and show that changes in the activity and possibly localization of transporters, along with the timing of sugar metabolism play a critical role in the sweet phenotype.ConclusionsThe high level of genomic similarity between sweet and grain sorghum reflects their historical relatedness, rather than their current phenotypic differences, but we find key changes in signaling molecules and transcriptional regulators that represent new candidates for understanding and improving sugar metabolism in this important crop.

Donor–acceptor (D–A) structure with charge transfer (CT) effect is widely utilized in the construction of organic luminescent materials. The adjustment of their CT effect and related luminescent property usually relies on the changes of molecular structure with different D or A moieties, which will lead to some uncertainties in structure‐property relationship. With the aim to explore the clear inherent luminescent mechanism for D–A molecule with CT effect, it is ideal that the regulation of intramolecular charge transfer in one same D–A molecule can be realized. Accordingly, three D–A type organic phosphorescence luminogens are designed and synthesized. Once being doped into different polymer matrixes, disparate charge transfer effects and related room temperature phosphorescence (RTP) properties can be achieved for a single molecule. Subsequent experiments confirm that different distributions of molecules with locally excited (LE) state and CT state are mainly responsible for their distinct RTP behaviors, exhibiting the well‐clarified structure‐property relationship of D–A type phosphorescence luminogens. Furthermore, the transition from CT to LE state can even be realized through chemical reaction, leading to the activated RTP effect for practical applications. [ABSTRACT FROM AUTHOR]

Objective: To investigate the ability of radiological parameter canal bone ratio (CBR) to assess bone mineral density and to differentiate between patients with primary and multiple osteoporotic vertebral compression fracture (OVCF). Methods: A retrospective analysis was conducted on OVCF patients treated at our hospital. CBR was measured through full-spine x-rays. Patients were categorized into primary and multiple fracture groups. Receiver operating characteristic curve analysis and area under the curve (AUC) calculation were used to assess the ability of parameters to predict osteoporosis and multiple fractures. Predictors of T values were analyzed by multiple linear regression, and independent risk factors for multiple fractures were determined by multiple logistic regression analysis. Results: CBR showed a moderate negative correlation with dual-energy x-ray absorptiometry T values (r = -0.642, p < 0.01). Higher CBR (odds ratio [OR], -6.483; 95% confidence interval [CI], -8.234 to -4.732; p < 0.01) and lower body mass index (OR, 0.054; 95% CI, 0.023–0.086; p < 0.01) were independent risk factors for osteoporosis. Patients with multiple fractures had lower T values (mean ±standard deviation [SD]: -3.76 ±0.73 vs. -2.83 ±0.75, p < 0.01) and higher CBR (mean ±SD: 0.54 ±0.07 vs. 0.46 ±0.06, p < 0.01). CBR had an AUC of 0.819 in predicting multiple fractures with a threshold of 0.53. T values prediction had an AUC of 0.816 with a threshold of -3.45. CBR > 0.53 was an independent risk factor for multiple fractures (OR, 14.66; 95% CI, 4.97–43.22; p < 0.01). Conclusion: CBR is negatively correlated with bone mineral density (BMD) and can be a novel opportunistic BMD assessment method. It is a simple and effective measurement index for predicting multiple fractures, with predictive performance not inferior to T values. [ABSTRACT FROM AUTHOR]

At present, the living habits of college students are relatively poor, and the amount of exercise is reduced, leading to their physical fitness getting worse and worse. Therefore, people began to study the physical health monitoring of college students. Machine learning and high‐performance computing in medical applications provide technical support for intelligent medical technology. With the rapid development of computer network, human beings have entered the information and digital era, and sports health big data has become more and more popular. The combination of sports health big data and student physical health monitoring technology, in a sense, can realize automatic data processing through intelligent medical and sports health database and other information technologies, thus promoting the popularization of health monitoring technology. However, the current health monitoring equipment has many problems, such as complex collection, low accuracy and limited processing of health data. To solve this problem, this paper developed an Application (APP) based on sports health big data technology that can monitor multiple vital signs such as human heart rate and body temperature, and analyze the physical health of college students, so that college students can easily understand their health status in daily life, so as to promote the healthy development of students' physique and encourage them to actively participate in physical exercise. The experiment proved that, in the heart rate monitoring, when the speed is 6 km/h, the error rate of the college students' physical health monitoring APP designed in this paper is 8.15%. The average accuracy rate of student steps monitoring is 97.12%. This showed the accuracy and availability of the APP's monitoring function for human vital signs. It has certain application value and significance to help students improve their physical quality. [ABSTRACT FROM AUTHOR]