Relationship between sample size and classification accuracy for tree species identification based on leaf shape

Land cover area estimation is one of the obvious applications of remote sensing. This paper compared two categories of area estimators, the confusion matrix calibration and the regression estimators, under various sample sizes and classification accuracy. Based on the relationship between the error distribution and heterogeneity of remote sensing data, simulation procedure was conducted to acquire the specific and controllable accuracy classification and preserve a real land cover structure pattern. And two criteria, the average absolute relative bias and CV, were adopted to evaluate the performance of estimators. The results suggest that (1) these estimators are asymptotically unbiased and zero-dispersed as sample size and the classification accuracy increases; (2) the high-quality classification data play a positive role in estimation and it is more meaningful for the confusion matrix estimators; (3) the regression estimators are slightly superior to the confusion matrix calibration estimators under acceptable classification accuracy level based on two criteria adopted.

A simple method for Eucalyptus species discrimination: FTIR spectroscopy and machine learning

The accurate identification of forest tree species is important for forest resource management and investigation. Using single remote sensing data for tree species identification cannot quantify both vertical and horizontal structural characteristics of tree species, so the classification accuracy is limited. Therefore, this study explores the application value of combining airborne high-resolution multispectral imagery and LiDAR data to classify tree species in study areas of different altitudes. Three study areas with different altitudes in Muyu Town, Shennongjia Forest Area were selected. Based on the object-oriented method for image segmentation, multi-source remote sensing feature extraction was performed. The recursive feature elimination algorithm was used to filter out the feature variables that were optimal for classifying tree species in each altitude study area. Four machine learning algorithms, SVM, KNN, RF, and XGBoost, were combined to classify tree species at each altitude and evaluate the accuracy. The results show that the diversity of tree layers decreased with the altitude in the different study areas. The texture features and height features extracted from LiDAR data responded better to the forest community structure in the different study areas. Coniferous species showed better classification than broad-leaved species within the same study areas. The XGBoost classification algorithm showed the highest accuracy of 87.63% (kappa coefficient of 0.85), 88.24% (kappa coefficient of 0.86), and 84.03% (kappa coefficient of 0.81) for the three altitude study areas, respectively. The combination of multi-source remote sensing numbers with the feature filtering algorithm and the XGBoost algorithm enabled accurate forest tree species classification.

BackgroundInformation on reproductive success is crucial to understanding population dynamics but can be difficult to obtain, particularly for species that birth while denning. For grizzly (Ursus arctos) and polar bears (U. maritimus), den visits are impractical because of safety and logistical considerations. Reproduction is typically documented through direct observation, which can be difficult, costly, and often occurs long after den departure. Reproduction could be documented remotely, however, from post-denning movement data if discernable differences exist between females with and without cubs.MethodsWe trained support vector machines (SVMs) with eight variables derived from telemetry data of female grizzly (2000–2022) and polar bears (1985–2016) with or without cubs during seven periods with lengths ranging from 5 to 60 days starting at den departure. We assessed SVM classification accuracy by withholding two samples (one cub-present, one cub-absent), training SVMs with the remaining data, predicting classification of the withheld samples, and repeating this process for each sample combination. Additionally, we evaluated how classification accuracy for grizzly bears was influenced by sample size, length of the post-departure period, and frequency of standardized location estimates.ResultsAccuracy of predicting cub presence or absence was 87% for grizzly bears with only 5 days of post-departure data and increased to a maximum of 92% with 20 days of data. For polar bears, accuracy was 86% at 5 days post-departure and increased to a maximum of 93% at 50 days. Classification accuracy for grizzly bears increased from 76 to 90% when sample size increased from 10 to 30 bears while holding period length constant (30 days) but did not increase at larger sample sizes. When sample size was held constant, increasing the length of the post-departure period did not affect classification accuracy markedly.ConclusionPresence or absence of grizzly and polar bear cubs can be identified with high accuracy even when SVM models are trained with limited data. Detecting cub presence or absence remotely could improve estimates of reproductive success and litter survival, enhancing our understanding of factors affecting cub recruitment.

Laser-Induced Breakdown Spectroscopy (LIBS) combined with Artificial Intelligence (AI) offers a powerful method for analyzing and comparing spectral data. This study presents a comparative analysis of conventional and AI-developed methods for processing and interpreting LIBS data, especially in forensic applications, focusing on toner sample discrimination. We propose a novel AI-developed approach that combines normalization, interpolation, and peak detection techniques to simplify LIBS spectral analysis without user preprocessing and easily identify unique spectral features. This method was compared with conventional principal component analysis (PCA) and partial least squares discriminant analysis (PLS-DA), which are commonly used for LIBS data analysis. The AI-developed method demonstrated superior performance in discriminating between toner samples from various brands and models of printers and photocopiers. The quantitative evaluation of the performance of the AI-developed approach was performed using statistical analysis, including accuracy difference percentage, component-wise variance analysis, paired t-test, and cross-validation test. The results confirmed a significant improvement in accuracy with the AI-developed method compared to conventional approaches. This proposed work highlights the potential of AI in enhancing spectroscopic analysis for forensic applications, offering increased efficiency and accuracy in sample discrimination and classification. Additionally, it accelerates the analysis of LIBS data with no need for user preprocessing.

Accurate identification of tree species is essential to monitor forest resources for climate change mitigation, biodiversity, and forest certification schemes. However, differentiating among species using remotely sensed data is challenging due to the similarity of spectral features. Here, we show that accuracy of species identification can be improved by incorporating phenological information derived from a time-series of multisource remote sensing data. Using the Google Earth Engine platform, we obtained a 4-year (2019–2022) time series of satellite imagery covering multiple phenological periods in mixed-species forests. This dataset was processed using Savitzky-Golay filtering and first-order spectral differential transformation to identify five dominant tree species through the Forest-Evergreen and Deciduous Forest-Tree Species Hierarchical Classification System (FEDT) with the Random Forest (RF) machine learning algorithm. The integration of phenological data, spectral indices and differential transformations achieved an overall accuracy of 0.82 and kappa coefficient of 0.75, compared to an overall accuracy of 0.76 and kappa coefficient of 0.68, respectively, when using spectral indices alone. Our findings highlight the value of time-series phenological analysis for enhancing the accuracy of tree species identification, providing a scalable method for improved monitoring of forest resources at regional to global scales.

This article highlights the critical role of reliable dc-dc converter operation in ensuring the stability of power conversion systems, especially in extreme environments like underwater observation networks. Addressing the common challenge in dc-dc converter fault diagnosis—overlooking the periodic characteristics of electrical signals during feature extraction from large datasets—we propose an innovative diagnostic method that combines adaptive wavelet transform (AWT) and an enhanced classifier generative adversarial network (ECGAN). First, AWT dynamically selects and optimizes WBFs to accurately capture degradation features in the output voltage signals. Then, a compound rainbow convolution block (CRConvBlock) is used to enhance the time-frequency representation of these signals, integrating temporal and frequency information for improved feature extraction. Furthermore, the proposed ECGAN model is guided by an adaptive loss optimization framework (ALOF) that dynamically adjusts training weights to balance sample quality and classification accuracy. Experimental validation in four different circuits demonstrates the high accuracy and robustness of the method, highlighting its potential for practical engineering applications.

BACKGROUND Detection of distinct epigenetic features in circulating cell-free DNA (cfDNA) of liquid biopsy specimens (e.g. blood) provides an opportunity to diagnose and prognosticate central nervous system (CNS) tumors. Utilization of distinctive cell-type genome-wide DNA methylation patterns allows for development of machine-learning (ML) models with the ability to predict tumor diagnosis and prognosis, and remains widely unexplored in meningiomas using LB. METHODS We profiled the cfDNA methylome (EPIC array) in serum specimens from patients with meningiomas, other CNS tumors and nontumor conditions (n= 63, 190 and 18, respectively) and harnessed internal and external meningioma tissue methylome data. For diagnostic model development, we identified “meningioma specific” signatures through comparison of meningioma and non-meningioma serum specimens (Wilcoxon rank-sum test) which exhibited tissue methylation similarities named Meningioma- epigenetic Liquid Biopsy (MeLB), and were then employed to develop and cross-validate a Random-Forest derived “MeLB” score to discriminate meningiomas from the other conditions. To predict recurrence risk (RR), we classified a meningioma tissue cohort as ‘favorable’ or ‘unfavorable’ (low and high RR, respectively), using a validated ML outcome model and identified outcome-specific methylation markers with serum subgroup specificities used as input to train a Random-Forest to predict RR in serum-based specimens. RESULTS Prediction models based on meningioma-specific methylation markers detected in the serum presented a high accuracy in classification of samples as meningioma or not (Accuracy: 89.6%, Sensitivity: 80%, Specificity: 93.8%). The prognostic model using tissue-serum matching methylation markers was validated across an independent tissue-based cohort (Accuracy: 88%, Sensitivity: 86%, Specificity: 88%) and allowed for classification of serum samples according to RR. CONCLUSIONS Machine-learning models using DNA methylation markers identified in the serum can accurately diagnose and predict prognosis in patients with meningioma. After validation in an external cohort, these approaches may improve presurgical diagnosis and therapeutic management of patients with this type of tumor.

Forest inventory, monitoring, and assessment requires accurate tree species identification and mapping. Recent experiences with multispectral data from small fixed-wing and rotary blade unmanned aerial vehicles (UAVs) suggest a role for this technology in the emerging paradigm of enhanced forest inventory (EFI). In this paper, pixel-based and object-based image analysis (OBIA) methods were compared in UAV-based tree species classification of nine commercial tree species in mature eastern Ontario mixedwood forests. Unsupervised clustering and supervised classification of tree crown pixels yielded approximately 50%–60% classification accuracy overall; OBIA with image segmentation to delineate tree crowns and machine learning yielded up to 80% classification accuracy overall. Spectral response patterns and tree crown shape and geometric differences were interpreted in context of their ability to separate tree species of interest with these classification methods. Accuracy assessment was based on field-based forest inventory tree species identification. The paper provides a brief summary of future research issues that will influence the growth of this geomatics innovation in forest tree species classification and forest inventory.

Methylation is one of the major epigenetic processes pivotal to our understanding of carcinogenesis. It is now widely accepted that there is a relationship between DNA methylation, chromatin structure, and human malignancies. DNA methylation is potentially an important clinical marker in cancer molecular diagnostics. Understanding epigenetic modifications in their biological context involves several aspects of DNA methylation analysis. These aspects include the de novo discovery of differentially methylated genes, the analysis of methylation patterns, and the determination of differences in the degree of methylation. Here we present a previously uncharacterized method for high-throughput DNA methylation analysis that utilizes MALDI-TOF mass spectrometry (MS) analysis of base-specifically cleaved amplification products. We use the IGF2/H19 region to show that a single base-specific cleavage reaction is sufficient to discover methylation sites and to determine methylation ratios within a selected target region. A combination of cleavage reactions enables the complete evaluation of all relevant aspects of DNA methylation, with most CpGs represented in multiple reactions. We successfully applied this technology under high-throughput conditions to quantitatively assess methylation differences between normal and neoplastic lung cancer tissue samples from 48 patients in 47 genes and demonstrate that the quantitative methylation results allow accurate classification of samples according to their histopathology.

Bee pollen may be contaminated with pyrrolizidine alkaloids (PAs) and their N-oxides (PANOs), which are mainly detected by liquid chromatography coupled to tandem mass spectrometry (LC–MS/MS), even though the use of fast near-infrared (NIR) spectroscopy is an ongoing alternative. Therefore, the main challenge of this study was to assess the feasibility of both a lab-stationary (Foss) and a portable (Polispec) NIR spectrometer in 60 dehydrated bee pollen samples. After an ANOVA-feature selection of the most informative NIR spectral data, canonical discriminant analysis (CDA) was performed to distinguish three quantitative PA/PANO classes (µg/kg): < LOQ (0.4), low; 0.4–400, moderate; > 400, high. According to the LC–MS/MS analysis, 77% of the samples were contaminated with PAs/PANOs and the sum content of the 17 target analytes was higher than 400 µg/kg in 28% of the samples. CDA was carried out on a pool of 18 (Foss) and 22 (Polispec) selected spectral variables and allowed accurate classification of samples from the low class as confirmed by the high values of Matthews correlation coefficient (≥ 0.91) for both NIR spectrometers. Leave-one-out cross-validation highlighted precise recognition of samples characterised by a high PA/PANO content with a low misclassification rate (0.02) as false negatives. The most informative wavelengths were within the < 1000, 1000–1660 and > 2400 nm regions for Foss and > 1500 nm for Polispec that could be associated with cyclic amines, and epoxide chemical structures of PAs/PANOs. In sum, both lab-stationary and portable NIR systems are reliable and fast techniques for detecting PA/PANO contamination in bee pollen.

Evaluation of near infrared spectroscopy for direct detection of common microplastics contamination in camel Milk powder.

Traditional taxonomy based on morphology has often failed in accurate species identification owing to the occurrence of cryptic species, which are reproductively isolated but morphologically identical. Molecular data have thus been used to complement morphology in species identification. The sexual advertisement calls in several groups of acoustically communicating animals are species-specific and can thus complement molecular data as non-invasive tools for identification. Several statistical tools and automated identifier algorithms have been used to investigate the efficiency of acoustic signals in species identification. Despite a plethora of such methods, there is a general lack of knowledge regarding the appropriate usage of these methods in specific taxa. In this study, we investigated the performance of two commonly used statistical methods, discriminant function analysis (DFA) and cluster analysis, in identification and classification based on acoustic signals of field cricket species belonging to the subfamily Gryllinae. Using a comparative approach we evaluated the optimal number of species and calling song characteristics for both the methods that lead to most accurate classification and identification. The accuracy of classification using DFA was high and was not affected by the number of taxa used. However, a constraint in using discriminant function analysis is the need for a priori classification of songs. Accuracy of classification using cluster analysis, which does not require a priori knowledge, was maximum for 6–7 taxa and decreased significantly when more than ten taxa were analysed together. We also investigated the efficacy of two novel derived acoustic features in improving the accuracy of identification. Our results show that DFA is a reliable statistical tool for species identification using acoustic signals. Our results also show that cluster analysis of acoustic signals in crickets works effectively for species classification and identification.

Studies of macro and microscopic wood anatomy as an auxiliary tool in taxonomy, particularly in the identification of forest tree species, have been used for decades. Comparative analysis of the wood anatomical structure allows the identification of unique morphological characteristics of the orders, families, genera, and in many cases, species, increasing the range of characters used in the botanical description and identification of tree species by classical taxonomy. In this work, techniques were used for comparative analysis and description of the anatomical structure of macro and microscopic wood tree species, that occur naturally in the Amazon rainforest in the Santarm, Moj Paragominas and Tailndia cities, in Par State. The aims of the study were (i) to characterize the macro and microscopic anatomical wood structure, (ii) identify the characteristics of taxonomic value through a comparative analysis of the wood anatomical structure, (iii) prepare the keys for species identification by macroscopic and microscopic wood anatomical structure. Were analyzed 100 wood samples from 64 tree species collected in 2004, in the project Dendrogene/Embrapa Amazonia Oriental. The surface of wood samples was polished with sandpaper and knives, according to their macro description; histological wood sections were prepared with razor sliding microtome, for microscopic wood description. The results allowed for the comparative analysis of the anatomical wood structure, to produce the macroscopic identification keys to 49 species of 51 genera. The genera Hymenolobium and Virola species with 4 and 2, respectively, could not be separated. The results from comparative wood anatomy showed evidence of 25 tree species with high potential for dendrochronological studies in order to distinct tree-ring presence (10 species) and barely distinct tree-rings (15 species). The application of wood anatomy, as an important tool with the potential of forest species taxonomy is discussed in this paper.

Morphological similarities among plant genera and species pose challenges in accurate identification, but leaf epidermis, with its diverse anatomical characteristics, serves as a valuable tool for differentiation. The study focused on the taxonomic significance of leaf epidermal features for the identification of various Ficus species. Leaves from different taxa were collected and preserved in 70% ethanol, and their epidermis carefully peeled off for microscopic examination. The staining process with safranin enhanced the visibility of various features. The observed anatomical characteristics included epidermal cell area, guard cell shape, stomatal size, stomatal shape, stomatal types, stomatal density, subsidiary cell shape, and types of trichomes. Significant variations were found in epidermal cell area, stomatal size, and stomatal density, etc. Clustering analysis categorized the species into four distinct groups, revealing associations and commonalities. Stomatal characteristics, subsidiary cell shapes, and trichomes presence are documented, which provide a detailed characterization. Positive and negative parameter-to-parameter correlations offered insights into trait interactions. The findings of this study contribute valuable information for the taxonomic identification of Ficus species, emphasizing the importance of leaf epidermal features in distinguishing among closely related taxa. This research enhances our understanding of plant diversity and provides a practical approach for accurate plant identification based on microscopic observations of leaf epidermal characteristics.