Search Results (17)

Many people admire the Japanese language and culture, but mastering the language’s writing system, particularly handwritten kanji, presents a significant challenge. Furthermore, translating historical manuscripts containing archaic or rare kanji requires specialized expertise. To address this, we designed a new model for handwritten kanji recognition based on the concept of cross-language transfer learning using a Preact ResNet-18 architecture. The model was pretrained in a Chinese dataset and subsequently fine-tuned in a Japanese dataset. We also adapted and evaluated two fine-tuning strategies: unfreezing only the last layer and unfreezing all the layers during fine-tuning. During the implementation of our training algorithms, we trained a model with the CASIA-HWDB dataset with handwritten Chinese characters and used its weights to initialize models that were fine-tuned with a Kuzushiji-Kanji dataset that consists of Japanese handwritten kanji. We investigated the effectiveness of the developed model when solving a multiclass classification task for three subsets with the one hundred fifty, two hundred, and three hundred most-sampled classes and showed an improvement in the recognition accuracy and an enhancement in a number of recognizable kanji with the proposed model compared to those of the existing methods. Our best model achieved 97.94% accuracy for 150 kanji, exceeding the previous SOTA result by 1.51%, while our best model for 300 kanji achieved 97.62% accuracy (exceeding the 150-kanji SOTA accuracy by 1.19% while doubling the class count). This confirms the effectiveness of our proposed model and establishes new benchmarks in handwritten kanji recognition, both in terms of accuracy and the number of recognizable kanji. Full article

Bioinformatics is a rapidly evolving field charged with cataloging, disseminating, and analyzing biological data. Bioinformatics started with genomics, but while genomics focuses more narrowly on the genes comprising a genome, bioinformatics now encompasses a much broader range of omics technologies. Overcoming barriers of scale and effort that plagued earlier sequencing methods, bioinformatics adopted an ambitious strategy involving high-throughput and highly automated assays. However, as the list of omics technologies continues to grow, the field of bioinformatics has changed in two fundamental ways. Despite enormous success in expanding our understanding of the biological world, the failure of bulk methods to account for biologically important variability among cells of the same or different type has led to a major shift toward single-cell and spatially resolved omics methods, which attempt to disentangle the conflicting signals contained in heterogeneous samples by examining individual cells or cell clusters. The second major shift has been the attempt to integrate two or more different classes of omics data in a single multimodal analysis to identify patterns that bridge biological layers. For example, unraveling the cause of disease may reveal a metabolite deficiency caused by the failure of an enzyme to be phosphorylated because a gene is not expressed due to aberrant methylation as a result of a rare germline variant. Conclusions: There is a fine line between superficial understanding and analysis paralysis, but like a detective novel, multi-omics increasingly provides the clues we need, if only we are able to see them. Full article

Motivation: Chest X-ray (CXR) is a routine diagnostic X-ray examination for checking and screening various diseases. Automatically localizing and classifying diseases from CXR as a detection task is of much significance for subsequent diagnosis and treatment. Due to the fact that samples of some diseases are difficult to acquire, CXR detection datasets often present a long-tail distribution over different diseases. Objective: The detection performance of tail classes is very poor due to the limited number and diversity of samples in the training dataset and should be improved. Method: In this paper, motivated by a correspondence-based tracking system, we build a pipeline named RaTrack, leveraging a diffusion model to alleviate the tail class degradation problem by aligning the generation process of the tail to the head class. Then, the samples of rare classes are generated to extend the number and diversity of rare samples. In addition, we propose a filtering strategy to control the quality of the generated samples. Results: Extensive experiments on public datasets, Vindr-CXR and RSNA, demonstrate the effectiveness of the proposed method, especially for rare diseases. Full article

Hyperspectral images (HSIs) provide valuable spatial–spectral information for ground analysis. However, in few-shot (FS) scenarios, the limited availability of training samples poses significant challenges in capturing the sample distribution under diverse environmental conditions. Semi-supervised learning has shown promise in exploring the distribution of unlabeled samples through pseudo-labels. Nonetheless, FS HSI classification encounters the issue of high intra-class spectral variability and inter-class spectral similarity, which often lead to the diffusion of unreliable pseudo-labels during the iterative process. In this paper, we propose a simple yet effective progressive pseudo-label selection strategy that leverages the spatial–spectral consistency of HSI pixel samples. By leveraging spatially aligned ground materials as connected regions with the same semantic and similar spectrum, pseudo-labeled samples were selected based on round-wise confidence scores. Samples within both spatially and semantically connected regions of FS samples were assigned pseudo-labels and joined subsequent training rounds. Moreover, considering the spatial positions of FS samples that may appear in diverse patterns, to fully utilize unlabeled samples that fall outside the neighborhood of FS samples but still belong to certain connected regions, we designed a matching active learning approach for expert annotation based on the temporal confidence difference. We identified samples with the highest training value in specific regions, utilizing the consistency between predictive labels and expert labels to decide whether to include the region or the sample itself in the subsequent semi-supervised iteration. Experiments on both classic and more recent HSI datasets demonstrated that the proposed base model achieved SOTA performance even with extremely rare labeled samples. Moreover, the extended version with active learning further enhances performance by involving limited additional annotation. Full article

The performance of semantic segmentation in remote sensing, based on deep learning models, depends on the training data. A commonly encountered issue is the imbalanced long-tailed distribution of data, where the head classes contain the majority of samples while the tail classes have fewer samples. When training with long-tailed data, the head classes dominate the training process, resulting in poorer performance in the tail classes. To address this issue, various strategies have been proposed, such as resampling, reweighting, and transfer learning. However, common resampling methods suffer from overfitting to the tail classes while underfitting the head classes, and reweighting methods are limited in the extreme imbalanced case. Additionally, transfer learning tends to transfer patterns learned from the head classes to the tail classes without rigorously validating its generalizability. These methods often lack additional information to assist in the recognition of tail class objects, thus limiting performance improvements and constraining generalization ability. To tackle the abovementioned issues, a graph neural network based on the graph kernel principle is proposed for the first time. By leveraging the graph kernel, structural information for tail class objects is obtained, serving as additional contextual information beyond basic visual features. This method partially compensates for the imbalance between tail and head class object information without compromising the recognition accuracy of head classes objects. The experimental results demonstrate that this study effectively addresses the poor recognition performance of small and rare targets, partially alleviates the issue of spectral confusion, and enhances the model’s generalization ability. Full article

With the growth of data in the real world, datasets often encounter the problem of long-tailed distribution of class sample sizes. In long-tailed image recognition, existing solutions usually adopt a class rebalancing strategy, such as reweighting based on the effective sample size of each class, which leans towards common classes in terms of higher accuracy. However, increasing the accuracy of rare classes while maintaining the accuracy of common classes is the key to solving the problem of long-tailed image recognition. This research explores a direction that balances the accuracy of both common and rare classes simultaneously. Firstly, a two-stage training is adopted, motivated by the use of transfer learning to balance features of common and rare classes. Secondly, a balanced weight function called Balanced Focal Softmax (BFS) loss is proposed, which combines balanced softmax loss focusing on common classes with balanced focal loss focusing on rare classes to achieve dual balance in long-tailed image recognition. Subsequently, a Balanced Weight Learning Mechanism (BWLM) to further utilize the feature of weight decay is proposed, where the weight decay as the weight balancing technique for the BFS loss tends to make the model learn smaller balanced weights by punishing the larger weights. Through extensive experiments on five long-tailed image datasets, it proves that transferring the weights from the first stage to the second stage can alleviate the bias of the naive models toward common classes. The proposed BWLM not only balances the weights of common and rare classes, but also greatly improves the accuracy of long-tailed image recognition and outperforms many state-of-the-art algorithms. Full article

In this investigation, we outline the applications of a data mining technique known as Subgroup Discovery (SD) to the analysis of a sample size-limited metabolomics-based dataset. The SD technique utilized a supervised learning strategy, which lies midway between classificational and descriptive criteria, in which given the descriptive property of a dataset (i.e., the response target variable of interest), the primary objective was to discover subgroups with behaviours that are distinguishable from those of the complete set (albeit with a differential statistical distribution). These approaches have, for the first time, been successfully employed for the analysis of aromatic metabolite patterns within an NMR-based urinary dataset collected from a small cohort of patients with the lysosomal storage disorder Niemann–Pick class 1 (NPC1) disease (n = 12) and utilized to distinguish these from a larger number of heterozygous (parental) control participants. These subgroup discovery strategies discovered two different NPC1 disease-specific metabolically sequential rules which permitted the reliable identification of NPC1 patients; the first of these involved ‘normal’ (intermediate) urinary concentrations of xanthurenate, 4-aminobenzoate, hippurate and quinaldate, and disease-downregulated levels of nicotinate and trigonelline, whereas the second comprised ‘normal’ 4-aminobenzoate, indoxyl sulphate, hippurate, 3-methylhistidine and quinaldate concentrations, and again downregulated nicotinate and trigonelline levels. Correspondingly, a series of five subgroup rules were generated for the heterozygous carrier control group, and ‘biomarkers’ featured in these included low histidine, 1-methylnicotinamide and 4-aminobenzoate concentrations, together with ‘normal’ levels of hippurate, hypoxanthine, quinolinate and hypoxanthine. These significant disease group-specific rules were consistent with imbalances in the combined tryptophan–nicotinamide, tryptophan, kynurenine and tyrosine metabolic pathways, along with dysregulations in those featuring histidine, 3-methylhistidine and 4-hydroxybenzoate. In principle, the novel subgroup discovery approach employed here should also be readily applicable to solving metabolomics-type problems of this nature which feature rare disease classification groupings with only limited patient participant and sample sizes available. Full article

In reviewing a selection of recent case studies from our laboratory, we revealed some lessons learned and benefits accrued from the application of mass spectrometry (MS/MS) molecular networking in the field of marine sponge natural products. Molecular networking proved pivotal to our discovery of many new natural products and even new classes of natural product, some of which were opaque to alternate dereplication and prioritization strategies. Case studies included the discovery of: (i) trachycladindoles, an exceptionally rare class of bioactive indole alkaloid previously only known from a single southern Australia sample of Trachycladus laevispirulifer; (ii) dysidealactams, an unprecedented class of sesquiterpene glycinyl-lactam and glycinyl-imide from a Dysidea sp., a sponge genera often discounted as having been exhaustively studied; (iii) cacolides, an unprecedented family of sesterterpene α-methyl-γ-hydroxybutenolides from a Cacospongia sp., all too easily mischaracterized and deprioritized during dereplication as a well-known class of sponge sesterterpene tetronic acids; and (iv) thorectandrins, a new class of indole alkaloid which revealed unexpected insights into the chemical and biological properties of the aplysinopsins, one of the earliest and more extensively reported class of sponge natural products. Full article

Previously, proof-of-concept studies have demonstrated that rare-earth elements (REEs) can be preferentially extracted from coal fly ash (CFA) solids using a recyclable ionic liquid (IL), betainium bis(trifluoromethylsulfonyl)imide ([Hbet][Tf₂N]). When the suspension of aqueous solution—IL-CFA—is heated above 65 °C, the majority of REEs will separate from the bulk elements in the solids and partition to the IL phase. Acid stripping of the IL removes REEs and regenerates the IL for reuse in additional extraction cycles. The objective of this study is to showcase the applicability and effectiveness of the optimized method to recover REEs from various CFAs. Six CFA samples with different characteristics (feed coal basins, coal beds, and ash collecting points) and classifications (Class C and Class F) were examined. The process performance was evaluated for a broad range of elements (33 total), including 15 REEs, two actinides, six bulk elements, and 10 trace metals. Results confirmed good recovery of total REEs (ranging from 44% to 66% among the CFA samples) and the recovery process’ high selectivity of REEs over other bulk and trace elements. Sc, Y, Nd, Sm, Gd, Dy, and Yb consistently showed high leaching and partitioning into the IL phase, with an average recovery efficiency ranging from 53.8% to 66.2%, while the other REEs showed greater variability among the different CFA samples. Some amounts of Al and Th were co-extracted into the IL phase, while Fe co-extraction was successfully limited by chloride complexation and ascorbic acid reduction. These results indicated that the IL-based REE-CFA recovery method can maintain a high REE recovery efficiency across various types of CFA, therefore providing a promising sustainable REE recovery strategy for various coal ash wastes. Full article

This paper presents a focused investigation into real-time segmentation in unstructured environments, a crucial aspect for enabling autonomous navigation in off-road robots. To address this challenge, an improved variant of the DDRNet23-slim model is proposed, which includes a lightweight network architecture and reclassifies ten different categories, including drivable roads, trees, high vegetation, obstacles, and buildings, based on the RUGD dataset. The model’s design includes the integration of the semantic-aware normalization and semantic-aware whitening (SAN–SAW) module into the main network to improve generalization ability beyond the visible domain. The model’s segmentation accuracy is improved through the fusion of channel attention and spatial attention mechanisms in the low-resolution branch to enhance its ability to capture fine details in complex scenes. Additionally, to tackle the issue of category imbalance in unstructured scene datasets, a rare class sampling strategy (RCS) is employed to mitigate the negative impact of low segmentation accuracy for rare classes on the overall performance of the model. Experimental results demonstrate that the improved model achieves a significant 14% increase mIoU in the invisible domain, indicating its strong generalization ability. With a parameter count of only 5.79M, the model achieves mAcc of 85.21% and mIoU of 77.75%. The model has been successfully deployed on a a Jetson Xavier NX ROS robot and tested in both real and simulated orchard environments. Speed optimization using TensorRT increased the segmentation speed to 30.17 FPS. The proposed model strikes a desirable balance between inference speed and accuracy and has good domain migration ability, making it applicable in various domains such as forestry rescue and intelligent agricultural orchard harvesting. Full article

Rare-earth oxyhydride (ReO_xH_y) films are novel inorganic photochromic materials that have strong potential for applications in windows and optical sensors. Cations greatly influence many material properties and play an important role in the photochromic performance of ReO_xH_y. Here we propose a strategy for obtaining Gd_1−zY_zO_xH_y films (z = 1, 0.7, 0.5, 0.4, 0.35, 0.25, 0.15, 0) using one-step direct-current (DC) magnetron co-sputtering. Distinct from the mixed anion systems, such material would belong to the class of mixed anion and mixed cation materials. For Gd_1−zY_zO_xH_y films, different co-doping ratios can help tune the contrast ratio (that is, the difference between coloration and bleaching transmittance) and cycling degradation, which may be related to the lattice constant. X-ray diffraction (XRD) patterns show that the lattice constant increases from 5.38 Å for YO_xH_y to 5.51 Å, corresponding to Gd_0.75Y_0.25O_xH_y. The contrast ratio, in particular, can be enhanced to 37% from 6.3% by increasing the lattice constant, directly controlled by the co-sputtering power. When the lattice constant decreases, the surface morphology of the sample with the smallest lattice constant is essentially unchanged by testing in air with normal oxidation for 100 days, suggesting great improvement in environment durability. However, the crystal structure cannot be overly compressed, and co-sputtering with Cr gives black opaque films without photochromic properties. Moreover, because the atomic mass of different rare earth elements is different, the critical pressure p* (films deposited at p < p* remain metallic dihydrides) is different, and the preparation window is enlarged. Our work provides insights into innovative photochromic materials that can help to achieve commercial production and application. Full article

Breast cancer is the most common disease among women, with over $2.1$ million new diagnoses each year worldwide. About $30\%$ of patients initially presenting with early stage disease have a recurrence of cancer within 10 years. Predicting who will have a recurrence and who will not remains challenging, with consequent implications for associated treatment. Artificial intelligence strategies that can predict the risk of recurrence of breast cancer could help breast cancer clinicians avoid ineffective overtreatment. Despite its significance, most breast cancer recurrence datasets are insufficiently large, not publicly available, or imbalanced, making these studies more difficult. This systematic review investigates the role of artificial intelligence in the prediction of breast cancer recurrence. We summarise common techniques, features, training and testing methodologies, metrics, and discuss current challenges relating to implementation in clinical practice. We systematically reviewed works published between 1 January 2011 and 1 November 2021 using the methodology of Kitchenham and Charter. We leveraged Springer, Google Scholar, PubMed, and IEEE search engines. This review found three areas that require further work. First, there is no agreement on artificial intelligence methodologies, feature predictors, or assessment metrics. Second, issues such as sampling strategies, missing data, and class imbalance problems are rarely addressed or discussed. Third, representative datasets for breast cancer recurrence are scarce, which hinders model validation and deployment. We conclude that predicting breast cancer recurrence remains an open problem despite the use of artificial intelligence. Full article

Endemic species are especially vulnerable to biodiversity loss caused by isolation or habitat specificity, small population size, and anthropogenic factors. Endemic species biodiversity analysis has a critically important global value for the development of conservation strategies. The rare onion Allium ledebourianum is a narrow-lined endemic species, with natural populations located in the extreme climatic conditions of the Kazakh Altai. A. ledebourianum populations are decreasing everywhere due to anthropogenic impact, and therefore, this species requires preservation and protection. Conservation of this rare species is associated with monitoring studies to investigate the genetic diversity of natural populations. Fundamental components of eukaryote genome include multiple classes of interspersed repeats. Various PCR-based DNA fingerprinting methods are used to detect chromosomal changes related to recombination processes of these interspersed elements. These methods are based on interspersed repeat sequences and are an effective approach for assessing the biological diversity of plants and their variability. We applied DNA profiling approaches based on conservative sequences of interspersed repeats to assess the genetic diversity of natural A. ledebourianum populations located in the territory of Kazakhstan Altai. The analysis of natural A. ledebourianum populations, carried out using the DNA profiling approach, allowed the effective differentiation of the populations and assessment of their genetic diversity. We used conservative sequences of tRNA primer binding sites (PBS) of the long-terminal repeat (LTR) retrotransposons as PCR primers. Amplification using the three most effective PBS primers generated 628 PCR amplicons, with an average of 209 amplicons. The average polymorphism level varied from 34% to 40% for all studied samples. Resolution analysis of the PBS primers showed all of them to have high or medium polymorphism levels, which varied from 0.763 to 0.965. Results of the molecular analysis of variance showed that the general biodiversity of A. ledebourianum populations is due to interpopulation (67%) and intrapopulation (33%) differences. The revealed genetic diversity was higher in the most distant population of A. ledebourianum LD64, located on the Sarymsakty ridge of Southern Altai. This is the first genetic diversity study of the endemic species A. ledebourianum using DNA profiling approaches. This work allowed us to collect new genetic data on the structure of A. ledebourianum populations in the Altai for subsequent development of preservation strategies to enhance the reproduction of this relict species. The results will be useful for the conservation and exploitation of this species, serving as the basis for further studies of its evolution and ecology. Full article

Rare cancers are identified as those with an annual incidence of fewer than 6 per 100,000 persons and includes both epithelial and stromal tumors from different anatomical areas. The advancement of analytical methods has produced an accurate molecular characterization of most human cancers, suggesting a “molecular classification” that has allowed the establishment of increasingly personalized therapeutic strategies. However, the limited availability of rare cancer samples has resulted in very few therapeutic options for these tumors, often leading to poor prognosis. Long non coding RNAs (lncRNAs) are a class of non-coding RNAs mostly involved in tumor progression and drug response. In particular, the lncRNA HOX transcript antisense RNA (HOTAIR) represents an emergent diagnostic, prognostic and predictive biomarker in many human cancers. The aim of this review is to highlight the role of HOTAIR in rare cancers, proposing it as a new biomarker usable in the management of these tumors. Full article

Due to spatial inhomogeneity of land-cover types and spectral confusions among them, land-cover maps suffer from misclassification errors. While much research has focused on improving image classification by re-processing source images with more advanced algorithms and/or using images of finer resolution, there is rarely any systematic work on re-processing existing maps to increase their accuracy. We propose refining existing maps to achieve accuracy gains by exploring and utilizing relationships between reference data, which are often already available or can be collected, and map data. For this, we make novel use of canonical correspondence analysis (CCA) to analyze reference-map class co-occurrences to facilitate probabilistic re-classification of map classes in CCA ordination space, a synthesized feature space constrained by map class occurrence patterns. Experiments using GlobeLand30 land-cover (2010) over Wuhan, China were carried out using reference sample data collected previously for accuracy assessment in the same area. Reference sample data were stratified by map classes and their spatial heterogeneity. To examine effects of model-training sample size on refinements, three subset samples (360, 720, and 1480 pixels) were selected from a pool of 3000 sample pixels (the full training sample). Logistic regression modeling was employed as a baseline method for comparisons. Performance evaluation was based on a test sample of 1020 pixels using a strict and relaxed definitions of agreement between reference classification and map classification, resulting in measures of types I and II, respectively. It was found that the CCA-based method is more accurate than logistic regression in general. With increasing sample sizes, refinements generally lead to greater accuracy gains. Heterogeneous sub-strata usually see greater accuracy gains than in homogeneous sub-strata. It was also revealed that accuracy gains in specific strata (map classes and sub-strata) are related to strata refinability. Regarding CCA-based refinements, a relatively small sample of 360 pixels achieved a 3% gain in both overall accuracy (OA) and F_0.01 score (II). By using a selective strategy in which only refinable strata of cultivated land and forest are included in refinement, accuracy gains are further increased, with 5–11% gains in users’ accuracies (UAs) (II) and 4–10% gains in F_0.01 scores (II). In conclusion, on condition of refinability, map refinement is well worth pursuing, as it increases accuracy of existing maps, extends utility of reference data, facilitates uncertainty-informed map representation, and enhances our understanding about relationships between reference data and map data and about their synthesis. Full article