Labeling Ratio Research Articles

ETFT: Equiangular Tight Frame Transformer for Imbalanced Semantic Segmentation.

Semantic segmentation often suffers from class imbalance, where the label ratio for each class in the dataset is not uniform. Recent studies have addressed the issue of class imbalance in semantic segmentation by leveraging the neural collapse phenomenon in conjunction with an Equiangular Tight Frame (ETF). While the use of ETF aids in enhancing the discriminability of minor classes, class correlation is another crucial factor that must be taken into account. However, managing the balance between class correlation and discrimination through neural collapse remains challenging, as these properties inherently conflict with one another. Moreover, this control is established during the training stage, resulting in a fixed classifier. There is no guarantee that this classifier will consistently perform well with different input images. To address this problem, we propose an Equiangular Tight Frame Transformer (ETFT), a transformer-based model that jointly processes the features and classifier using ETF structure, and dynamically generates the classifier as a function of the input for imbalanced semantic segmentation. Specifically, the classifier initialized with the ETF structure is jointly processed with the input patch tokens during the attention process. As a result, the transformed patch tokens, aided by the ETF structure, achieve discriminability between classes while preserving contextual correlation. The classifier, initially structured as an ETF, is adjusted to incorporate the correlation information, benefiting from the attention mechanism. Furthermore, the learned classifier is combined with the fixed ETF classifier, leveraging the advantages of both. Extensive experiments demonstrate that the proposed method outperforms state-of-the-art methods for imbalanced semantic segmentation on both the ADE20K and Cityscapes datasets.

Mathematical Modeling Unveils Optimization Strategies for Targeted Radionuclide Therapy of Blood Cancers.

Mathematical modeling yields general principles for optimization of TRT in mouse models of multiple myeloma that can be extrapolated to other cancer models and clinical settings.

Decreased solubility and increased adsorptivity of a biotinylated humanized anti-cocaine mAb

Biotinylation of proteins, including antibodies, is a very useful and important modification for a variety of biochemical characterizations, including anti-drug antibody (ADA) assays used to detect antibodies raised against therapeutic antibodies. We assessed different degrees of biotin labeling of an anti-cocaine mAb currently under development for treating cocaine use disorder. We noted that higher levels of biotin labeling dramatically decreased mAb solubility, and increased the tendency to bind to surfaces, complicating characterization of the biotinylated antibody. Specifically, in phosphate buffered saline, labeling stoichiometries of more than about 3 biotin/mAb resulted in decreased recoveries due to increased binding to surfaces and decreased mAb solubility. Native gel agarose electrophoresis, differential scanning fluorimetry, and isothermal titration calorimetry all demonstrated changes in the mAb which became more pronounced above a labeling ratio of 3 biotin/mAb. At 3.0 biotin/mAb, there were minimal changes in solubility, adsorptivity, exposure of hydrophobic dye-binding sites, heat stability, and cocaine binding, in stark contrast to labeling with 5.6 biotin/mAb. Thus, the degree of biotinylation should be kept at about 3 biotin/mAb to maintain antigen binding and general structure, solubility, and stability of this mAb, a finding which may be important for other similar mAbs currently in use or under development.

A Low-Measurement-Cost-Based Multi-Strategy Hyperspectral Image Classification Scheme.

The cost of hyperspectral image (HSI) classification primarily stems from the annotation of image pixels. In real-world classification scenarios, the measurement and annotation process is both time-consuming and labor-intensive. Therefore, reducing the number of labeled pixels while maintaining classification accuracy is a key research focus in HSI classification. This paper introduces a multi-strategy triple network classifier (MSTNC) to address the issue of limited labeled data in HSI classification by improving learning strategies. First, we use the contrast learning strategy to design a lightweight triple network classifier (TNC) with low sample dependence. Due to the construction of triple sample pairs, the number of labeled samples can be increased, which is beneficial for extracting intra-class and inter-class features of pixels. Second, an active learning strategy is used to label the most valuable pixels, improving the quality of the labeled data. To address the difficulty of sampling effectively under extremely limited labeling budgets, we propose a new feature-mixed active learning (FMAL) method to query valuable samples. Fine-tuning is then used to help the MSTNC learn a more comprehensive feature distribution, reducing the model's dependence on accuracy when querying samples. Therefore, the sample quality is improved. Finally, we propose an innovative dual-threshold pseudo-active learning (DSPAL) strategy, filtering out pseudo-label samples with both high confidence and uncertainty. Extending the training set without increasing the labeling cost further improves the classification accuracy of the model. Extensive experiments are conducted on three benchmark HSI datasets. Across various labeling ratios, the MSTNC outperforms several state-of-the-art methods. In particular, under extreme small-sample conditions (five samples per class), the overall accuracy reaches 82.97% (IP), 87.94% (PU), and 86.57% (WHU).

Diagnosis of high‐speed railway ballastless track arching based on unsupervised learning framework

AbstractVehicle‐mounted detection methods have been widely applied in the maintenance of high‐speed railways (HSRs), providing feasibility for diagnosing ballastless track arching. However, applying detection data faces several key limitations: (1) The threshold mostly requires manual setting, making recognition accuracy highly subjective; (2) the extensive workload of manual inspections makes it challenging to label detection data, hindering the application of supervised learning approaches. To address these problems, this paper utilizes the longitudinal level irregularity data obtained from vehicle‐mounted detection, employing the concept of unsupervised learning for dimensionality reduction, combined with clustering algorithms and minimal label fine‐tuning, to design two frameworks: the fully unsupervised framework (FUF) and the few‐shot fine‐tuned framework (FFF). Experiments on dynamic detection data from a Chinese HSR line were conducted, comparing the performance of data dimensionality reduction, clustering, and classification under different strategy combinations. The results show that the improved variational autoencoder significantly enhances the performance of the encoder in dimensionality reduction, facilitating better feature extraction; the FUF achieves effective clustering outcomes without any labeled samples and its adjusted rand index score exceeded 0.8, showcasing its robustness and applicability in scenarios with no prior annotations; the FFF requires only a small number of labeled samples (labeling ratio of 5%) and achieves excellent performance, with metrics such as accuracy exceeding 0.85, thus greatly reducing the reliance on labeled data. This study offers a novel method for solving engineering issues with limited labeled data, providing an efficient solution for identifying track arching defects and advancing railway infrastructure monitoring.

Semi-Supervised Medical Image Classification with Pseudo Labels Using Coalition Similarity Training

The development of medical image classification models necessitates a substantial number of labeled images for model training. In real-world scenarios, sample sizes are typically limited and labeled samples often constitute only a small portion of the dataset. This paper aims to investigate a collaborative similarity learning strategy that optimizes pseudo-labels to enhance model accuracy and expedite its convergence, known as the joint similarity learning framework. By integrating semantic similarity and instance similarity, the pseudo-labels are mutually refined to ensure their quality during initial training. Furthermore, the similarity score is utilized as a weight to guide samples away from misclassification predictions during the classification process. To enhance the model’s generalization ability, an adaptive consistency constraint is introduced into the loss function to improve performance on untrained datasets. The model achieved a satisfactory accuracy of 93.65% at 80% labeling ratio, comparable to supervised learning methods’ performance. Even with very low labeling ratio (e.g., 5%), the model still attained an accuracy of 74.28%. Comparison with other techniques such as Mean Teacher and FixMatch revealed that our approach significantly outperforms them in medical image classification tasks through improving accuracy by approximately 2%, demonstrating this framework’s leadership in medical image classification.

Prediction of rhinitis with class imbalance based on heterogeneous ensemble learning

Common clinical rhinitis is characterized by different types of cases and class imbalance. Its prediction belongs to multiple output classification. Low recognition rate and poor generalization performance often occur for minority class. Therefore, we propose a novel integrated classification model, ARF-OOBEE, which transforms the multi-output classification to multi-label classification and multi-class classification. The multi-label classifier automatically adjusts the number and depth of integrated forest learners according to the imbalance ratio of single class label in a subset. It can effectively reduce the impact of class imbalance on classification and improve prediction performance of both majority or minority class concurrently. Also, we build a multi-class classification based on out-of-bag Extra-Tree to accomplish finer classification for the predicted labels. In addition, we calculate the feature importance for rhinitis on the grounds of the purity of nodes in decision-making tree inside Random Forest and study the correlation between rhinitis features. We conduct 12 folds cross-validation experiments on 461 cases of clinical rhinitis. The outcomes show that the evaluation indicators of ARF-OOBEE, such as Sensitivity, Specificity, Accuracy, F1-Score, AUC, and G-Mean are 74.9%,86.5%,92.0%,78.3%,95.3%, and 79.9%, respectively. In comparison to the other methods, ARF-OOBEE has better evaluation indicator and is more effective for the early clinical diagnosis of rhinitis.

Conversational AI: An Explication of Few-Shot Learning Problem in Transformers-Based Chatbot Systems

With the recent advancements in conversational artificial intelligence (AI), the practical applications of chatbots have risen significantly in diverse domains such as healthcare, education, e-government, customer support systems, social platforms, and entertainment. The chatbot converses with humans in natural language and responds to their queries with precise and relevant answers. The relevant literature presents several methods for intent classification and slot mapping for chatbot question answering. However, the existing chatbot architectures still suffer from few-shot learning problem (imbalance ratio of class labels over samples) and ineffective dialog management to retain the context and slots mapping. This study aims to introduce the architecture of chatbot by focusing few-shot learning problem and context management in dialog-based conversations. First, to mitigate the few-shot learning problem with chatbots, we propose a novel hybrid intent and slots transformers (HIST) model. The HIST chatbot architecture utilizes transformers and self-attention mechanisms along with bigated recurrent unit and combines conditional random field algorithms for intent classification and slot extraction. Second, to address dialog management, we introduce a hybrid interaction strategy for slots mapping and effective conversational context management. To validate the proposed model’s effectiveness, comprehensive empirical analysis is carried out using three benchmark datasets including airline travel information system (ATIS), banking77, and conversational language interface for natural conversation 150 (CLINC150). The results show that HIST outperforms against the state-of-the-art existing methods with a clear margin and obtained an accuracy of 94.89% and 96.17% for intent classification and slots extraction, respectively. The empirical results confirm the effectiveness of the HIST chatbot for resolving the few-shot learning problem with effective dialog management in chatbot systems.

Dual-Perspective Knowledge Enrichment for Semi-supervised 3D Object Detection

Semi-supervised 3D object detection is a promising yet under-explored direction to reduce data annotation costs, especially for cluttered indoor scenes. A few prior works, such as SESS and 3DIoUMatch, attempt to solve this task by utilizing a teacher model to generate pseudo-labels for unlabeled samples. However, the availability of unlabeled samples in the 3D domain is relatively limited compared to its 2D counterpart due to the greater effort required to collect 3D data. Moreover, the loose consistency regularization in SESS and restricted pseudo-label selection strategy in 3DIoUMatch lead to either low-quality supervision or a limited amount of pseudo labels. To address these issues, we present a novel Dual-Perspective Knowledge Enrichment approach named DPKE for semi-supervised 3D object detection. Our DPKE enriches the knowledge of limited training data, particularly unlabeled data, from two perspectives: data-perspective and feature-perspective. Specifically, from the data-perspective, we propose a class-probabilistic data augmentation method that augments the input data with additional instances based on the varying distribution of class probabilities. Our DPKE achieves feature-perspective knowledge enrichment by designing a geometry-aware feature matching method that regularizes feature-level similarity between object proposals from the student and teacher models. Extensive experiments on the two benchmark datasets demonstrate that our DPKE achieves superior performance over existing state-of-the-art approaches under various label ratio conditions. The source code and models will be made available to the public.

Quantitative N‐ or C‐Terminal Labelling of Proteins with Unactivated Peptides by Use of Sortases and a d‐Aminopeptidase

AbstractQuantitative and selective labelling of proteins is widely used in both academic and industrial laboratories, and catalytic labelling of proteins using transpeptidases, such as sortases, has proved to be a popular strategy for such selective modification. A major challenge for this class of enzymes is that the majority of procedures require an excess of the labelling reagent or, alternatively, activated substrates rather than simple commercially sourced peptides. We report the use of a coupled enzyme strategy which enables quantitative N‐ and C‐terminal labelling of proteins using unactivated labelling peptides. The use of an aminopeptidase in conjunction with a transpeptidase allows sequence‐specific degradation of the peptide by‐product, shifting the equilibrium to favor product formation, which greatly enhances the reaction efficiency. Subsequent optimisation of the reaction allows N‐terminal labelling of proteins using essentially equimolar ratios of peptide label to protein and C‐terminal labelling with only a small excess. Minimizing the amount of substrate required for quantitative labelling has the potential to improve industrial processes and facilitate the use of transpeptidation as a method for protein labelling.

Quantitative N- or C-Terminal Labelling of Proteins with Unactivated Peptides by Use of Sortases and a d-Aminopeptidase.

Quantitative and selective labelling of proteins is widely used in both academic and industrial laboratories, and catalytic labelling of proteins using transpeptidases, such as sortases, has proved to be a popular strategy for such selective modification. A major challenge for this class of enzymes is that the majority of procedures require an excess of the labelling reagent or, alternatively, activated substrates rather than simple commercially sourced peptides. We report the use of a coupled enzyme strategy which enables quantitative N- and C-terminal labelling of proteins using unactivated labelling peptides. The use of an aminopeptidase in conjunction with a transpeptidase allows sequence-specific degradation of the peptide by-product, shifting the equilibrium to favor product formation, which greatly enhances the reaction efficiency. Subsequent optimisation of the reaction allows N-terminal labelling of proteins using essentially equimolar ratios of peptide label to protein and C-terminal labelling with only a small excess. Minimizing the amount of substrate required for quantitative labelling has the potential to improve industrial processes and facilitate the use of transpeptidation as a method for protein labelling.

A SILAC-based accurate quantification of shrimp allergen tropomyosin in complex food matrices using UPLC-MS/MS

The carryover of trace allergens in complex food matrices poses challenges for detection techniques. Here, we demonstrate an accurate UPLC-MS/MS quantification assay for the shrimp allergen tropomyosin with a full-length isotope-labelled recombinant tropomyosin (TM-I) internal standard in complex food matrices. The TM-I, expressed based on the SILAC technique, exhibited a high isotope labelling ratio (>99%), purity, and alignment with the natural sequence. This method determined the tropomyosin ranging from 0.2 to 100 ng/mL. Mean recoveries ranged from 89 to 116%, with intra- and inter-day RSDs below 12%, for three signature peptides across three types of commercially processed food matrices. The limits of quantitation were 1 μg/g in pop food and sauce, and 10 μg/g in surimi product, respectively. This study supports the use of recombinant full-length isotope-labelled proteins rather than stable-isotope labelling peptides as internal standards to achieve more accurate quantitation of food allergens as the digestion error is corrected.

Metabolic signatures of thymomas: potential biomarkers and treatment targets.

A study of tumour metabolic reprogramming has revealed disease biomarkers and avenues for therapeutic intervention. Metabolic reprogramming in thymoma is currently understudied and largely unknown. This study utilized metabolomics and isotope tracing with 13C-glucose to metabolically investigate thymomas, adjacent thymic tissue and benign thymic lesions. From 2017 to 2021, 20 patients with a suspected thymoma were recruited to this prospective Institutional Review Board approved clinical trial. At the time of surgery, 11 patients were infused with 13C-glucose, a stable, non-radioactive tracer which reports the flow of carbon through metabolic pathways. Samples were analysed by mass spectrometry to measure the abundance of >200 metabolites.13C enrichment was measured in patients who received 13C-glucose infusions. Histological analysis showed that 9 patients had thymomas of diverse subtypes and 11 patients had benign cysts. In our metabolomic analysis, thymomas could be distinguished from both adjacent thymus tissue and benign lesions by metabolite abundances. Metabolites in pyrimidine biosynthesis and glycerophospholipid metabolism were differentially expressed across these tissues.13C-glucose infusions revealed differential labelling patterns in thymoma compared to benign cysts and normal thymus tissue. The lactate/3PG labelling ratio, a metabolic marker in aggressive lung tumours correlated with lactate uptake, was increased in thymomas (1.579) compared to normal thymus (0.945) and benign masses (0.807) (thymic tissue versus tumour P = 0.021, tumour versus benign P = 0.013). We report metabolic biomarkers, including differential 13C labelling of metabolites from central metabolism, that distinguish thymomas from benign tissues. Altered glucose and lactate metabolism warrant further investigation and may provide novel therapeutic targets for thymoma.

Mix-Teaching: A Simple, Unified and Effective Semi-Supervised Learning Framework for Monocular 3D Object Detection

Semi-supervised learning (SSL) has promising potential for improving model performance using both labelled and unlabelled data. Since recovering 3D information from 2D images is an ill-posed problem, the current state-of-the-art methods of monocular 3D object detection (Mono3D) have relatively low precision and recall, making semi-supervised learning for Mono3D tasks challenging and understudied. In this work, we propose a unified and effective semi-supervised learning framework called Mix-Teaching that can be applied to most monocular 3D object detectors. Based on the idea of decomposition and recombination, unlabelled samples are firstly decomposed into collections of image patches with high-quality predictions and collections of background images containing no objects. The student model is then trained on the mixed images containing dense instances with high-quality pseudo-labels generated by the recombination operation. In addition, we propose an uncertainty-based filter to distinguish high-quality pseudo-labels from noisy predictions during the decomposition process. As results in KITTI and nuScenes benchmarks, Mix-Teaching consistently improves MonoFlex and GUPNet by significant margins under various labeling ratios. Our method achieves around +6.34% <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AP</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3<i>D</i></sub> improvement against the GUPNet on the validation set when using only 10% labelled data. Using the full training set and the additional 38K raw images from KITTI, it can further improve the MonoFlex by +4.65% absolute improvement on <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AP</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3<i>D</i></sub> for car detection, reaching 18.54% <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">AP</i> <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">3<i>D</i></sub> , which ranks the 1st place among all monocular based methods on the KITTI test leaderboard.

Evolutionary measures and their correlations with the performance of cross‐version defect prediction for object‐oriented projects

AbstractCross‐version defect prediction (CVDP) for evolutionary projects has attracted much attention from researchers in recent years. For multiple versions of an object‐oriented project, the degree of evolution (e.g., the degree of class change) between successive versions can reflect the differences between versions, which could affect the performance of CVDP. Therefore, how to measure the degree of evolution between successive versions and explore the correlations with the performance of CVDP are very important for software defect prediction. Based on the successive versions of evolutionary projects, this paper proposes six evolutionary measures from three aspects of class change, metric change, and label change, including the Ratio of New Classes (RNC), the Ratio of Deleted Classes (RDC), the Average Ratio of Metric Change (ARMC), the Ratio of Label Changed Classes (RLCC), the Ratio of Unchanged Classes (RUC), and the Ratio of Interference Classes (RIC). An empirical study was conducted on 40 versions of 11 object‐oriented projects from the PROMISE repository. Precision, Recall, F‐measure, and AUC were used as the performance indicators. Three correlation approaches (Pearson, Spearman, and Kendall) are applied to show the correlations between evolutionary measures and the performance of CVDP. The statistical results show that RNC, RDC, and RUC show no correlation with four performance indicators. ARMC shows weak or medium positive correlations with Recall and F‐measure. RLCC and RIC show very strong or strong negative correlations with Recall and F‐measure. The results indicate that the correlations between the proposed evolutionary measures and the performance of CVDP are different, which can guide the training set selection of CVDP.

Ensemble selector mixed with pareto optimality to feature reduction

Among various tasks of feature reduction, how to search qualified features and then construct the corresponding reduct which satisfies a given constraint is a hot topic. Constraint is the pre-defined condition in feature reduction, it rules two crucial aspects: 1) when the process of searching can be terminated to construct reduct; 2) whether a feature/set of features should be added into reduct pool or not. Obviously, such a task of feature reduction is constraint-dependent. However, some limitations may naturally emerge: 1) constraint-independent features are ignored; 2) fixed constraint hinders the diverse evaluations of features; 3) selected features based on single constraint are powerless to data perturbation. Therefore, an Ensembler Mixed with Pareto Optimality (EmPo) is developed in this study. Firstly, the principle of Restricted Pareto Optimality is proposed to identify constraint-independent features before performing constraint related searching, which indicates a two-stages strategy to feature reduction. Naturally, diverse evaluations of features are achieved in such a process. Secondly, a data perturbation w.r.t. either sample or feature aspect is employed to obtain multiple reductions of features. The objective is to further improve the stability of the classification results based on those multiple reducts. It should also be emphasized that EmPo is a general framework, and most existing approaches to feature reduction can be embedded into it to further improve their performances. Finally, the effectiveness of our EmPo is validated over 20 datasets within 4 different ratios of noisy label.

Use of Magnetotactic Bacteria as an MRI Contrast Agent for In Vivo Tracking of Adoptively Transferred Immune Cells.

In vivo immune cell tracking using MRI can be a valuable tool for studying the mechanisms underlying successful cancer therapies. Current cell labeling methods using superparamagnetic iron oxide (SPIO) lack the persistence to track the fate and location of transplanted cells long-term. Magnetospirillum magneticum is a commercially available, iron-producing bacterium that can be taken up by and live harmoniously within mammalian cells as magneto-endosymbionts (MEs). MEs have shown promise as labeling agents for in vivo stem and cancer cell tracking but have yet to be evaluated in immune cells. This pilot study examined ME labeling in myeloid-derived suppressor cells (MDSCs), cytotoxic T lymphocytes (CTLs), and dendritic cells (DCs) and its effects on cell purity, function, and MRI contrast. MDSCs, CTLs, and DCs were incubated with MEs at various ME labeling ratios (MLR), and various biological metrics and iron uptake were assessed. For in vivo imaging, MDSCs were labeled overnight with either MEs or SPIO (Molday ION Rhodamine B) and injected into C3 tumor-bearing mice via tail vein injection 24days post-implant and scanned daily with MRI for 1week to assess cellular quantification. Following incubations, MDSCs contained > 0.6pg Fe/cell. CTLs achieved Fe loading of < 0.5pg/cell, and DCs achieved Fe loading of ~ 1.4pg/cell. The suppressive functionality of MDSCs at 1000 MLR was not affected by ME labeling but was affected at 2000 MLR. Markers of CTL dysfunction were not markedly affected by ME labeling nor were DC markers. In vivo data demonstrated that the MDSCs labeled with MEs generated sufficient contrast to be detectable using TurboSPI, similar to SPIO-labeled cells. Cells can be labeled with sufficient numbers of MEs to be detectable with MRI without compromising cell viability. Care must be taken at higher concentrations of MEs, which may affect some cell types' functional activity and/or morphology. Immune cells with minimal phagocytic behavior have much lower iron content per cell after incubation with MEs vs SPIO; however, MEs can successfully be used as a contrast agent for phagocytic immune cells.

ISL-GKFDA: An incomplete supervision manifold learning framework based on propagation graph of data

To improve manifold classification results using unlabeled data with high dimensional redundancy, we propose an incomplete supervision manifold learning framework. In this framework, Euclidean distance and Euclidean neighborhood are replaced by manifold distance and manifold neighborhood respectively. The pseudo label of unlabeled data is generated by constructing label propagation network in manifold neighborhood. By adding the local manifold neighborhood structure into KFDA (Kernel Fisherś Discriminant Analysis) as regularization term, the local manifold structure is stressed. Subsequently, we propose the ISL-GKFDA (Incomplete Supervision Learning (ISL) with propagation Graph of data in Kernel Fisher Discriminant Analysis) algorithm. We validate ISL-GKFDA by comparing it with FDA (Fisher's Discrimination Analysis), KFDA (Kernel FDA), ISL-FDA (ISL version FDA), ISL-KFDA (ISL version KFDA), SSDA (semi-supervised discriminant analysis based on manifold distance), and KSSDA (Kernel SSDA). In typical manifold datasets, ISL-GKFDA can more accurately capture the boundaries of binary classification; and visually in handwritten digital dataset, ISL-GKFDA can separate three data types more clearly. Finally, we use the UCI database and compare multiple algorithms. When measured by accuracy, the proposed ISL-GKFDA reaches a maximum in at least three datasets of five when labeling ratio is 5%, 15% and 30%, respectively. When measured by F1, ISL-GKFDA reaches the maximum in all datasets.

Improvement and Assessment of Convolutional Neural Network for Tree Species Identification Based on Bark Characteristics

Efficient tree species identification is of great importance in forest inventory and management. As the textural properties of tree barks vary less notably as a result of seasonal change than other tree organs, they are more suitable for the identification of tree species using deep learning models. In this study, we adopted the ConvNeXt convolutional neural network to identify 33 tree species using the BarkNetV2 dataset, compared the classification accuracy values of different tree species, and performed visual analysis of the network’s visual features. The results show the following trends: (1) the pre-trained network weights exhibit up to 97.61% classification accuracy for the test set, indicating that the network has high accuracy; (2) the classification accuracy values of more than half of the tree species can reach 98%, while the confidence level of correct identification (probability ratio of true labels) of tree species images is relatively high; and (3) there is a strong correlation between the network’s visual attractiveness and the tree bark’s biological characteristics, which share similarities with humans’ organization of tree species. The method suggested in this study has the potential to increase the efficiency of tree species identification in forest resources surveys and is of considerable value in forest management.

Quality of amoxicillin/clavulanic acid oral formulations for intended veterinary use in the UK, Malaysia, Serbia and Thailand.

Amoxicillin/clavulanate is the most commonly used oral antimicrobial drug in companion animals. The objective of the study was to detect types and frequency of deficits in the quality of veterinary oral formulations of amoxicillin/clavulanate in various countries. In a prospective study with purposive sampling, amoxicillin/clavulanate tablet formulations for canine use were collected in four countries (wholesalers or veterinary practice) and shipped to a central bioanalytical laboratory. Twenty-four samples were collected from the UK (nine), Malaysia (nine), Serbia (four) and Thailand (two), yielding 18 different formulations (10 veterinary). Packaging inspection, tablet disintegration and content assay were conducted (validated high-performance liquid chromatography with ultra-violet detection); content was acceptable when within the 90% to 120% pre-specified range (US Pharmacopeia). Secondary packaging was present for 13 of 24 samples and primary packaging integrity was verified for all but one sample. Amoxicillin trihydrate/potassium clavulanate label ratio was 4:1, except for three formulations (2:1). Tablet dose strength ranged from 250 to 625 mg. All formulations contained both analytes. For amoxicillin, two of 24 samples were out of specification with 72.8% (Malaysia) and 82.3% (Thailand) of labelled content. For clavulanate, four of 24 samples were out of specification with 46.9% (Serbia), 79.0% (UK), 84.3% (Serbia) and 86.5% (Thailand) of labelled content. One formulation (Thailand) failed for both analytes. Antimicrobial formulations of substandard quality have negative consequences for efficacy in patients and potentially promote antimicrobial resistance. There was evidence of substandard formulations in all countries, not only for amoxicillin but especially for clavulanate; this could compromise equitable access to acceptable quality essential veterinary medicines worldwide.