A robust deep attention dense convolutional neural network for plant leaf disease identification and classification from smart phone captured real world images

Plant diseases cause significant food loss and hence economic loss around the globe. Therefore, automatic plant disease identification is a primary task to take proper medications for controlling the spread of the diseases. Large variety of plants species and their dissimilar phytopathological symptoms call for the implementation of supervised machine learning techniques for efficient and reliable disease identification and classification. With the development of deep learning strategies, convolutional neural network (CNN) has paved its way for classification of multiple plant diseases by extracting rich features. However, several characteristics of the input images especially captured in real world environment, viz. complex or indistinguishable background, presence of multiple leaves with the diseased leaf, small lesion area, solemnly affect the robustness and accuracy of the CNN modules. Available strategies usually applied standard CNN architectures on the images captured in the laboratory environment and very few have considered practical in-field leaf images for their studies. However, those studies are limited with very limited number of plant species. Therefore, there is need of a robust CNN module which can successfully recognize and classify the dissimilar leaf health conditions of non-identical plants from the in-field RGB images. To achieve the above goal, an attention dense learning (ADL) mechanism is proposed in this article by merging mixed sigmoid attention learning with the basic dense learning process of deep CNN. The basic dense learning process derives new features at higher layer considering all lower layer features and that provides fast and efficient training process. Further, the attention learning process amplifies the learning ability of the dense block by discriminating the meaningful lesion portions of the images from the background areas. Other than adding an extra layer for attention learning, in the proposed ADL block the output features from higher layer dense learning are used as an attention mask to the lower layers. For an effective and fast classification process, five ADL blocks are stacked to build a new CNN architecture named DADCNN-5 for obtaining classification robustness and higher testing accuracy. Initially, the proposed DADCNN-5 module is applied on publicly available extended PlantVillage dataset to classify 38 different health conditions of 14 plant species from 54,305 images. Classification accuracy of 99.93% proves that the proposed CNN module can be used for successful leaf disease identification. Further, the efficacy of the DADCNN-5 model is checked after performing stringent experiments on a new real world plant leaf database, created by the authors. The new leaf database contains 10,851 real-world RGB leaf images of 17 plant species for classifying their 44 distinguished health conditions. Experimental outcomes reveal that the proposed DADCNN-5 outperforms the existing machine learning and standard CNN architectures, and achieved 97.33% accuracy. The obtained sensitivity, specificity and false positive rate values are 96.57%, 99.94% and 0.063% respectively. The module takes approximately 3235 min for training process and achieves 99.86% of training accuracy. Visualization of Class activation mapping (CAM) depicts that DADCNN-5 is able to learn distinguishable features from semantically important regions (i.e. lesion regions) on the leaves. Further, the robustness of the DADCNN-5 is established after experimenting with augmented and noise contaminated images of the practical database.     

BLeafNet: A Bonferroni mean operator based fusion of CNN models for plant identification using leaf image classification

Plants, the only natural source of oxygen, are the most important resources for every species in the world. A proper identification of plants is important for different fields. The observation of leaf characteristics is a popular method as leaves are easily available for examination. Researchers are increasingly applying image processing techniques for the identification of plants based on leaf images. In this paper, we have proposed a leaf image classification model, called BLeafNet, for plant identification, where the concept of deep learning is combined with Bonferroni fusion learning. Initially, we have designed five classification models, using ResNet-50 architecture, where five different inputs are separately used in the models. The inputs are the five variants of the leaf grayscale images, RGB, and three individual channels of RGB - red, green, and blue. For fusion of the five ResNet-50 outputs, we have used the Bonferroni mean operator as it expresses better connectivity among the confidence scores, and it also obtains better results than the individual models. We have also proposed a two-tier training method for properly training the end-to-end model. To evaluate the proposed model, we have used the Malayakew dataset, collected at the Royal Botanic Gardens in New England, which is a very challenging dataset as many leaves from different species have a very similar appearance. Besides, the proposed method is evaluated using the Leafsnap and the Flavia datasets. The obtained results on both the datasets confirm the superiority of the model as it outperforms the results achieved by many state-of-the-art models.     

Automated classification of urinary stones based on microcomputed tomography images using convolutional neural network

PurposeThe classification of urinary stones is important prior to treatment because the treatments depend on three types of urinary stones, i.e., calcium, uric acid, and mixture stones. We have developed an automatic approach for the classification of urinary stones into the three types based on microcomputed tomography (micro-CT) images using a convolutional neural network (CNN).Materials and methodsThirty urinary stones from different patients were scanned in vitro using micro-CT (pixel size: 14.96 μm; slice thickness: 15 μm); a total of 2,430 images (micro-CT slices) were produced. The slices (227 × 227 pixels) were classified into the three categories based on their energy dispersive X-ray (EDX) spectra obtained via scanning electron microscopy (SEM). The images of urinary stones from each category were divided into three parts; 66%, 17%, and 17% of the dataset were assigned to the training, validation, and test datasets, respectively. The CNN model with 15 layers was assessed based on validation accuracy for the optimization of hyperparameters such as batch size, learning rate, and number of epochs with different optimizers. Then, the model with the optimized hyperparameters was evaluated for the test dataset to obtain classification accuracy and error.ResultsThe validation accuracy of the developed approach with CNN with optimized hyperparameters was 0.9852. The trained CNN model achieved a test accuracy of 0.9959 with a classification error of 1.2%.ConclusionsThe proposed automated CNN-based approach could successfully classify urinary stones into three types, namely calcium, uric acid, and mixture stones, using micro-CT images.     

The leaf economics spectrum’s morning coffee: plant size-dependent changes in leaf traits and reproductive onset in a perennial tree crop

Background and AimsSize-dependent changes in plant traits are an important source of intraspecific trait variation. However, there are few studies that have tested if leaf trait co-variation and/or trade-offs follow a within-genotype leaf economics spectrum (LES) related to plant size and reproductive onset. To our knowledge, there are no studies on any plant species that have tested whether or not the shape of a within-genotype LES that describes how traits covary across whole plant sizes, is the same as the shape of a within-genotype LES that represents environmentally driven trait plasticity.MethodsWe quantified size-dependent variation in eight leaf traits in a single coffee genotype (Coffea arabica var. Caturra) in managed agroecosystems with different environmental conditions (light and fertilization treatments), and evaluated these patterns with respect to reproductive onset. We also evaluated if trait covariation along a within-genotype plant-size LES differed from a within-genotype environmental LES defined with trait data from coffee growing in different environmental conditions.Key ResultsLeaf economics traits related to resource acquisition – maximum photosynthetic rates (A) and mass-based leaf nitrogen (N) concentrations – declined linearly with plant size. Structural traits – leaf mass, leaf thickness, and leaf mass per unit area (LMA) – and leaf area increased with plant size beyond reproductive onset, then declined in larger plants. Three primary LES traits (mass-based A, leaf N and LMA) covaried across a within-genotype plant-size LES, with plants moving towards the ‘resource-conserving’ end of the LES as they grow larger; in coffee these patterns were nearly identical to a within-genotype environmental LES.ConclusionsOur results demonstrate that a plant-size LES exists within a single genotype. Our findings indicate that in managed agroecosystems where resource availability is high the role of reproductive onset in driving within-genotype trait variability, and the strength of covariation and trade-offs among LES traits, are less pronounced compared with plants in natural systems. The consistency in trait covariation in coffee along both plant-size and environmental LES axes indicates strong constraints on leaf form and function that exist within plant genotypes.     

Delimiting cryptic morphological variation among human malaria vector species using convolutional neural networks

Deep learning is a powerful approach for distinguishing classes of images, and there is a growing interest in applying these methods to delimit species, particularly in the identification of mosquito vectors. Visual identification of mosquito species is the foundation of mosquito-borne disease surveillance and management, but can be hindered by cryptic morphological variation in mosquito vector species complexes such as the malaria-transmitting Anopheles gambiae complex. We sought to apply Convolutional Neural Networks (CNNs) to images of mosquitoes as a proof-of-concept to determine the feasibility of automatic classification of mosquito sex, genus, species, and strains using whole-body, 2D images of mosquitoes. We introduce a library of 1, 709 images of adult mosquitoes collected from 16 colonies of mosquito vector species and strains originating from five geographic regions, with 4 cryptic species not readily distinguishable morphologically even by trained medical entomologists. We present a methodology for image processing, data augmentation, and training and validation of a CNN. Our best CNN configuration achieved high prediction accuracies of 96.96% for species identification and 98.48% for sex. Our results demonstrate that CNNs can delimit species with cryptic morphological variation, 2 strains of a single species, and specimens from a single colony stored using two different methods. We present visualizations of the CNN feature space and predictions for interpretation of our results, and we further discuss applications of our findings for future applications in malaria mosquito surveillance.     

Deep learning-based multi-task prediction system for plant disease and species detection

The manual prediction of plant species and plant diseases is expensive, time-consuming, and requires expertise that is not always available. Automated approaches, including machine learning and deep learning, are increasingly being applied to surmount these challenges. For this, accurate models are needed to provide reliable predictions and guide the decision-making process. So far, these two problems have been addressed separately, and likewise, separate models have been developed for each of these two problems, but considering that plant species and plant disease prediction are often related tasks, they can be considered together. We therefore propose and validate a novel approach based on the multi-task learning strategy, using shared representations between these related tasks, because they perform better than individual models. We apply a multi-input network that uses raw images and transferred deep features extracted from a pre-trained deep model to predict each plant's type and disease. We develop an end-to-end multi-task model that carries out more than one learning task at a time and combines the Convolutional Neural Network (CNN) features and transferred features. We then evaluate this model using public datasets. The results of our experiments demonstrated that this Multi-Input Multi-Task Neural Network model increases efficiency and yields faster learning for similar detection tasks.     

Detection method for Convallaria keiskei colonies in Hokkaido,Japan, by combining CNN and FCM using UAV-based remote sensing data

The Convallaria keiskei, a plant species indigenous to Japan, is on the verge of extinction. In the past, they have been manually protected and managed. Unmanned aerial vehicles (UAVs), which are already being applied in various fields, such as agriculture, surveying, and logistics, can be applied to automate this task. Image processing and machine learning techniques applied on images obtained from UAVs can automate Convallaria keiskei classification, help to estimate the increase in colony numbers, and reduce the detection cost. In a previous study, a flower number estimation method that combines image processing and a convolutional neural network (CNN) was proposed. However, leaf regions similar to flower regions were misidentified as flower regions, and the accuracy was reduced. Therefore, in this study, a method was investigated to reduce the number of false positives by excluding areas similar to the flower regions. Specifically, a novel detection method combining image processing, CNN, and fuzzy c-means is proposed. To validate the proposed method, it was compared with the previous method as well as the method in which k-means was used instead of fuzzy c-means. All results were evaluated using flower distribution maps marked by field workers. The proposed method improved the F-measure by up to 22.0% compared with the previous method. Application of the proposed method to orthorectified images facilitates the understanding of flower populations over a wide range of areas, which can contribute to the conservation and management of the species.     

Accurate Prediction of Transposon-Derived piRNAs by Integrating Various Sequential and Physicochemical Features

BackgroundPiwi-interacting RNA (piRNA) is the largest class of small non-coding RNA molecules. The transposon-derived piRNA prediction can enrich the research contents of small ncRNAs as well as help to further understand generation mechanism of gamete.MethodsIn this paper, we attempt to differentiate transposon-derived piRNAs from non-piRNAs based on their sequential and physicochemical features by using machine learning methods. We explore six sequence-derived features, i.e. spectrum profile, mismatch profile, subsequence profile, position-specific scoring matrix, pseudo dinucleotide composition and local structure-sequence triplet elements, and systematically evaluate their performances for transposon-derived piRNA prediction. Finally, we consider two approaches: direct combination and ensemble learning to integrate useful features and achieve high-accuracy prediction models.ResultsWe construct three datasets, covering three species: Human, Mouse and Drosophila, and evaluate the performances of prediction models by 10-fold cross validation. In the computational experiments, direct combination models achieve AUC of 0.917, 0.922 and 0.992 on Human, Mouse and Drosophila, respectively; ensemble learning models achieve AUC of 0.922, 0.926 and 0.994 on the three datasets.ConclusionsCompared with other state-of-the-art methods, our methods can lead to better performances. In conclusion, the proposed methods are promising for the transposon-derived piRNA prediction. The source codes and datasets are available in S1 File.     

A New Coastal Species of Limonium (Plumbaginaceae) from Southeastern Spain

A new species, Limonium scopulorum (Plumbaginaceae), is described from the maritime cliffs of Alicante province, Spain (southeastern part of the Iberian Peninsula). This new taxon belongs to the group of Limonium delicatulum, which is highly diversified in the Mediterranean territories of the Iberian Peninsula. It is related to the Balearic species L. biflorum, though several leaf, floral and chromosomal features warrant its easy recognition. Molecular divergence, shown in previous studies, also supports its separate treatment. Major affinities and differences with other related taxa are discussed. Moreover, morphological, ecological, chorological, biogeographic features of the new taxon are discussed, and its conservation status is reported. An identification key is provided for Spanish taxa of the Limonium delicatulum group.     

Effects of phylogeny,leaf traits,and the altitudinal distribution of host plants on herbivore assemblages on congeneric Acer species

Historical, niche-based, and stochastic processes have been proposed as the mechanisms that drive community assembly. In plant–herbivore systems, these processes can correspond to phylogeny, leaf traits, and the distribution of host plants, respectively. Although patterns of herbivore assemblages among plant species have been repeatedly examined, the effects of these factors among co-occurring congeneric host plant species have rarely been studied. Our aim was to reveal the process of community assembly for herbivores by investigating the effects of phylogeny, leaf traits, and the altitudinal distribution of closely related host plants of the genus Acer. We sampled leaf functional traits for 30 Acer species in Japan. Using a newly constructed phylogeny, we determined that three of the six measured leaf traits (leaf thickness, C/N ratio, and condensed tannin content) showed a phylogenetic signal. In a field study, we sampled herbivore communities on 14 Acer species within an elevation gradient and examined relationships between herbivore assemblages and host plants. We found that herbivore assemblages were significantly correlated with phylogeny, leaf traits, phylogenetic signals, and the altitudinal distribution of host plants. Our results indicate that the interaction between historical and current ecological processes shapes herbivore community assemblages.     

Multi-Feature Fusion Method for Identifying Carotid Artery Vulnerable Plaque

PurposeVulnerable plaque of carotid atherosclerosis is prone to rupture, which can easily lead to acute cardiovascular and cerebrovascular accidents. Accurate identification of the vulnerable plaque is a challenging task, especially on limited datasets.MethodsThis paper proposes a multi-feature fusion method to identify high-risk plaque, in which three types of features are combined, i.e. global features of carotid ultrasound images, echo features of regions of interests (ROI) and expert knowledge from ultrasound reports. Due to the fusion of three types of features, more critical features for identifying high-risk plaque are included in the feature set. Therefore, better performance can be achieved even on limited datasets.ResultsFrom testing all combinations of three types of features, the results showed that the accuracy of using all three types of features is the highest. The experiments also showed that the performance of the proposed method is better than other plaque classification methods and classical Convolutional Neural Networks (CNNs) on the Plaque dataset.ConclusionThe proposed method helped to build a more complete feature set so that the machine learning models could identify vulnerable plaque more accurately even on datasets with poor quality and small scale.     

Leaf traits and performance vary with plant age and water availability in Artemisia californica

Background and AimsLeaf functional traits are strongly tied to growth strategies and ecological processes across species, but few efforts have linked intraspecific trait variation to performance across ontogenetic and environmental gradients. Plants are believed to shift towards more resource-conservative traits in stressful environments and as they age. However, uncertainty as to how intraspecific trait variation aligns with plant age and performance in the context of environmental variation may limit our ability to use traits to infer ecological processes at larger scales.MethodsWe measured leaf physiological and morphological traits, canopy volume and flowering effort for Artemisia californica (California sagebrush), a dominant shrub species in the coastal sage scrub community, under conditions of 50, 100 and 150 % ambient precipitation for 3 years.Key ResultsPlant age was a stronger driver of variation in traits and performance than water availability. Older plants demonstrated trait values consistent with a more conservative resource-use strategy, and trait values were less sensitive to drought. Several trait correlations were consistent across years and treatments; for example, plants with high photosynthetic rates tended to have high stomatal conductance, leaf nitrogen concentration and light-use efficiency. However, the trade-off between leaf construction and leaf nitrogen evident in older plants was absent for first-year plants. While few traits correlated with plant growth and flowering effort, we observed a positive correlation between leaf mass per area and performance in some groups of older plants.ConclusionsOverall, our results suggest that trait sensitivity to the environment is most visible during earlier stages of development, after which intraspecific trait variation and relationships may stabilize. While plant age plays a major role in intraspecific trait variation and sensitivity (and thus trait-based inferences), the direct influence of environment on growth and fecundity is just as critical to predicting plant performance in a changing environment.     

Osservazioni Sulla Maggiorana (Origanum Majorana L.) in Veste di Pianta Coltivata

Abstract

Stomatal features and ontogeny of stomata of 11 ornamental taxa of monocotyledonous families with Agavaceae (1 species), Amaryllidaceae (1 species), Araceae (3 species), Cannaceae (1 species), Commelinaceae (3 species), Liliaceae (1 species), and Musaceae (1 species) have been studied. Features like stomatal area, leaf area occupied by stomata and per cent leaf area occupied by stomata are reported for these taxa for the first time.     

DNA Barcoding of Rhodiola (Crassulaceae): A Case Study on a Group of Recently Diversified Medicinal Plants from the Qinghai-Tibetan Plateau

DNA barcoding, the identification of species using one or a few short standardized DNA sequences, is an important complement to traditional taxonomy. However, there are particular challenges for barcoding plants, especially for species with complex evolutionary histories. We herein evaluated the utility of five candidate sequences — rbcL, matK, trnH-psbA, trnL-F and the internal transcribed spacer (ITS) — for barcoding Rhodiola species, a group of high-altitude plants frequently used as adaptogens, hemostatics and tonics in traditional Tibetan medicine. Rhodiola was suggested to have diversified rapidly recently. The genus is thus a good model for testing DNA barcoding strategies for recently diversified medicinal plants. This study analyzed 189 accessions, representing 47 of the 55 recognized Rhodiola species in the Flora of China treatment. Based on intraspecific and interspecific divergence and degree of monophyly statistics, ITS was the best single-locus barcode, resolving 66% of the Rhodiola species. The core combination rbcL+matK resolved only 40.4% of them. Unsurprisingly, the combined use of all five loci provided the highest discrimination power, resolving 80.9% of the species. However, this is weaker than the discrimination power generally reported in barcoding studies of other plant taxa. The observed complications may be due to the recent diversification, incomplete lineage sorting and reticulate evolution of the genus. These processes are common features of numerous plant groups in the high-altitude regions of the Qinghai-Tibetan Plateau.     

Seasonal and between-sex differences in the diet of Black Grouse Tetrao tetrix

Capsule Dietary differences between sexes and seasons reflected diversity in plant availability and habitat preferences.

Aims To analyse Black Grouse diet through the identification of plant and invertebrate material found in the crop.

Methods Crops were dissected and the content categorized into plant species and components (i.e. leaf, stem, flower, fruit and seed). Multivariate analysis was used to test for seasonal, sex- or location-related differences in the diet.

Results Plant fragments accounted for 98% of the diet and included 53 plant species or taxa. Invertebrates made up the remaining 2%. Diet varied significantly between seasons and sexes. Both sexes ate more ericaceous shrubs in autumn and winter, with females eating more than males. The plant parts eaten varied seasonally. In summer, fruits, flowers and seeds were favoured over leaves, which dominated in winter.

Conclusions With few trees, birds were reliant upon Heather in autumn and winter.     

Root taxa identification in plant mixtures – current techniques and future challenges

Background

Studying root biomass, root system distribution and belowground interactions is essential for understanding the composition of plant communities, the impact of global change, and terrestrial biogeochemistry. Most soil samples and minirhizotron pictures hold roots of more than one species or plant individual. The identification of taxa by their roots would allow species-specific questions to be posed; information about root affiliation to plant individuals could be used to determine intra-specific competition.

Scope

Researchers need to be able to discern plant taxa by roots as well as to quantify abundances in mixed root samples. However, roots show less distinctive features that permit identification than aboveground organs. This review discusses the primary use of available methods, outlining applications, shortcomings and future developments.

Conclusion

Methods are either non-destructive, e.g. visual examination of root morphological criteria in situ, or require excavated and excised root samples. Among the destructive methods are anatomical keys, chemotaxonomic approaches and molecular markers. While some methods allow for discerning the root systems of individual plants, others can distinguish roots on the functional group or plant taxa level; methods such as IR spectroscopy and qPCR allow for quantifying the root biomass proportion of species without manual sorting.     

DNA Barcoding as an Effective Tool in Improving a Digital Plant Identification System: A Case Study for the Area of Mt. Valerio, Trieste (NE Italy)

Background

Identification keys are decision trees which require the observation of one or more morphological characters of an organism at each step of the process. While modern digital keys can overcome several constraints of classical paper-printed keys, their performance is not error-free. Moreover, identification cannot be always achieved when a specimen lacks some morphological features (i.e. because of season, incomplete development or miss-collecting). DNA barcoding was proven to have great potential in plant identification, while it can be ineffective with some closely related taxa, in which the relatively brief evolutionary distance did not produce differences in the core-barcode sequences.

Methodology/Principal Findings

In this paper, we investigated how the DNA barcoding can support the modern digital approaches to the identification of organisms, using as a case study a local flora, that of Mt. Valerio, a small hill near the centre of Trieste (NE Italy). The core barcode markers (plastidial rbcL and matK), plus the additional trnH-psbA region, were used to identify vascular plants specimens. The usefulness of DNA barcoding data in enhancing the performance of a digital identification key was tested on three independent simulated scenarios.

Conclusions/Significance

Our results show that the core barcode markers univocally identify most species of our local flora (96%). The trnH-psbA data improve the discriminating power of DNA barcoding among closely related plant taxa. In the multiparametric digital key, DNA barcoding data improves the identification success rate; in our simulation, DNA data overcame the absence of some morphological features, reaching a correct identification for 100% of the species. FRIDA, the software used to generate the digital key, has the potential to combine different data sources: we propose to use this feature to include molecular data as well, creating an integrated identification system for plant biodiversity surveys.