Threshold Selection Research Articles

U-Net transfer learning backbones for lesions segmentation in breast ultrasound images

<p>Breast ultrasound images are highly valuable for the early detection of breast cancer. However, the drawback of these images is low-quality resolution and the presence of speckle noise, which affects their interpretability and makes them radiologists’ expertise-dependent. As medical images, breast ultrasound datasets are scarce and imbalanced, and annotating them is tedious and time-consuming. Transfer learning, as a deep learning technique, can be used to overcome the dataset deficiency in available images. This paper presents the implementation of transfer learning U-Net backbones for the automatic segmentation of breast ultrasound lesions and implements a threshold selection mechanism to deliver optimal generalized segmentation results of breast tumors. The work uses the public breast ultrasound images (BUSI) dataset and implements ten state-of-theart candidate models as U-Net backbones. We have trained these models with a five-fold cross-validation technique on 630 images with benign and malignant cases. Five out of ten models showed good results, and the best U-Net backbone was found to be DenseNet121. It achieved an average Dice coefficient of 0.7370 and a sensitivity of 0.7255. The model’s robustness was also evaluated against normal cases, and the model accurately detected 72 out of 113 images, which is higher than the four best models.</p>

Event-driven online-learning using the CAR-FAC cochlea model

This work explores using a local spike-timing-dependent adaptation of thresholds and weights to learn to classify spectro-temporal representations of audio represented by neural spikes. We use the Cascade of Asymmetric Resonators with Fast Acting Compression (CAR-FAC) cochlea model and Leaky Integrated-and-Fire (LIF) neurons to generate the spikes from audio. This event stream is fed into the Optimised Deep Event-driven Spiking Architecture (ODESA) to learn spectro-temporal features in a hierarchical architecture. The cochlear events provide robust spectro-temporal representations of audio, and the ODESA performs online learning without needing error back-propagation or the calculation of gradients. Using simple local adaptive selection thresholds at each node, ODESA rapidly learns to appropriately allocate its neuronal resources at each layer in the hierarchy to learn features at different spatial and temporal scales. Information transmission throughout the system is event-based and all computing is performed asynchronously and online. We tested the approach on the TIDIGTIS benchmark and compare the performance with existing event-driven TIDIGTIS datasets and feature extraction and classification algorithms. With a three-layer ODESA, the proposed approach achieves 84.2% accuracy.

IDRM: Brain tumor image segmentation with boosted RIME optimization

Timely diagnosis of medical conditions can significantly mitigate the risks they pose to human life. Consequently, there is an urgent demand for an effective auxiliary model that assists physicians in accurately diagnosing medical conditions based on imaging data. While multi-threshold image segmentation models have garnered considerable attention due to their simplicity and ease of implementation, the selection of threshold combinations greatly influences the segmentation performance. Traditional optimization algorithms often require substantial time to address multi-threshold image segmentation problems, and their segmentation accuracy is frequently unsatisfactory. As a result, metaheuristic algorithms have been employed in this domain. However, several algorithms suffer from drawbacks such as premature convergence and inadequate exploration of the solution space when it comes to threshold selection. For instance, the recently proposed optimization algorithm RIME, inspired by the physical phenomenon of rime-ice, falls short in terms of avoiding local optima and fully exploring the solution space. Therefore, this study introduces an enhanced version of RIME, called IDRM, which incorporates an interactive mechanism and Gaussian diffusion strategy. The interactive mechanism facilitates information exchange among agents, enabling them to evolve towards more promising directions and increasing the likelihood of discovering the optimal solution. Additionally, the Gaussian diffusion strategy enhances the agents' local exploration capabilities and expands their search within the solution space, effectively preventing them from becoming trapped in local optima. Experimental results on 30 benchmark test functions demonstrate that IDRM exhibits favorable optimization performance across various optimization functions, showcasing its robustness and convergence properties. Furthermore, the algorithm is applied to select threshold combinations for brain tumor image segmentation, and the results are evaluated using metrics such as Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM). The overall findings consistently highlight the exceptional performance of this approach, further validating the effectiveness of IDRM in addressing image segmentation problems.

Prospective Trial of Using Imaging to Predict Pathologic Response and Clinical Outcomes in Locally Advanced Esophageal Cancer

Trimodality therapy with chemoradiation (CRT) followed by esophagectomy is the standard of care for locally advanced esophageal cancer. An unresolved question is whether pathologic complete response (pCR) can be assessed non-invasively for patients post-CRT. In this study, we assessed whether diffusion-weighted imaging (DWI) with MRI or PET can be used as predictors of pCR and other clinical outcomes after CRT. Patients were enrolled on a single-arm institutional trial (PA13-0380) assessing the role of imaging in predicting outcomes in potentially resectable esophageal patients undergoing trimodality therapy. All patients received neoadjuvant CRT, and 29 patients had subsequent surgery. DWI MRI and PET scans were obtained at baseline, 2 weeks after the start of CRT (interim) and 4 to 6 weeks after completion of CRT (follow up). Apparent diffusion coefficients (ADCs) were calculated based on DWI images. Circulating tumor DNA was obtained for 27 patients post-radiation using CAPP-Seq. Mann-Whitney tests compared imaging changes associated with pCR. Discrimination of pCR by imaging changes was quantified by received operating characteristics. Youden's index was applied to select optimal thresholds. Kaplan-Meier analysis was performed to assess differences in overall survival (OS) and progression-free survival (PFS) by changes in DWI, PET, and ctDNA parameters. Our cohort of 60 patients had a median follow up of 42.7 months, age of 65.4 yrs, and ECOG of 1 at completion of CRT. 90% were male, 58% had a history of smoking, and 85% were white. 83% had adenocarcinoma with the rest squamous cell carcinoma. Stages of the patients ranged from IIA to IIIB. All had moderately (47%) or poorly (53%) differentiated disease. All received 41.4-50.4 Gy in 1.8 Gy fractions with the majority receiving 50.4 Gy (95%). 29 patients underwent surgery after CRT of which 8 (27.6%) had pCR. Mean ΔADC from baseline to mid-treatment was most associated with pCR (AUC = 0.98, p<0.001) for patients undergoing surgery. Max ΔADC from baseline to first follow-up was most associated with OS (p = 0.002) and PFS (p<0.001) for the whole cohort. 27 patients had ctDNA analyzed after RT with the presence of ctDNA significantly associated with worse OS (HR = 0.12, p = 0.05) and PFS (HR = 0.10, p = 0.002). Combining ctDNA and max ΔADC generated a model that was more predictive of OS and PFS than either alone. We found that neither the PET parameters of TLG or SUV max at baseline or changes in these parameters from baseline to mid-treatment or first follow-up were as predictive as DWI. We show that changes in DWI is associated with pCR, OS, and PFS in resectable esophageal cancer patients undergoing CRT. DWI was more predictive than PET and a model combining DWI and ctDNA was the most predictive of clinical outcomes. This study shows the significant promise of using DWI in potentially guiding treatment decisions in esophageal cancer patients and will require validation in a larger cohort.

Comparison of Estimated Late Toxicities between IMRT and IMPT when Treating Retroperitoneal Sarcoma Preoperatively with Ultra-Hypofractionation

While preoperative radiation did not improve abdominal recurrence-free survival for retroperitoneal sarcoma (RPS) in the randomized STRASS trial, it did reduce rates of local recurrence. The risk of radiation-associated toxicity is substantial, with 77% of irradiated patients experiencing grade ≥3 lymphopenia in STRASS. Thus, finding methods to mitigate the issues of delaying surgery and irradiating normal tissue may provide a path towards affording the benefits of radiation while limiting its downside. One potential solution involves the use of hypofractionation to limit delay from radiation initiation to surgery, and proton therapy to limit dose to surrounding organs at risk (OARs). We conducted a dosimetric comparison of preoperative ultra-hypofractionated intensity-modulated photon radiotherapy (IMRT) and proton therapy (IMPT) for RPS, comparing estimated rates of late toxicity using published normal tissue complication probability (NTCP) models. Volumetric modulated arc therapy IMRT and IMPT plans were generated on 10 RPS patients previously treated with preoperative radiation. The prescription was 25 Gy radiobiological equivalent (GyE) to the clinical target volume (CTV) and 30 GyE to the margin-at-risk, all in five fractions. Proton doses were calculated using a radiobiological effective dose of 1.1. NTCPs were calculated for each OAR as a function of equivalent uniform dose. The ΔNTCP (difference in absolute NTCP between IMRT and IMPT plans) for each of the toxicity domains was calculated. Student T-tests were used to compare differences in dosimetric and NTCP outcomes. CTV coverage was met for all IMRT and IMPT plans with >99% of CTVs receiving ≥100% of prescription doses. The following endpoints were significantly lower with IMPT than IMRT: mean doses to liver, bone, and all analyzed genitourinary and gastrointestinal OARs; bowel, kidney, and bone V5-V20; stomach V15; liver V5; maximum doses to stomach, spinal canal, and body; and whole-body integral dose. No OAR endpoint was significantly higher with IMPT. The average ΔNTCP for grade 3 bowel ulceration/perforation and renal toxicity was 1.9% (p = .037) and 43.0% (p = .023), respectively, favoring IMPT. Using a model-based selection threshold of any ΔNTCP >10%, 50% (n = 5) of patients would be eligible for IMPT. IMPT maintained target coverage while significantly reducing dose to adjacent OARs and integral dose compared to IMRT. This translated to significantly lower risks of estimated late gastrointestinal and renal toxicities with IMPT. Further investigation is warranted to validate these findings and potential clinical benefit in the management of RPS. A prospective trial treating RPS with preoperative ultra-hypofractionated IMPT at our institution is currently being pursued (NCT05302570).

Improved pairs trading strategy using two-level reinforcement learning framework

Pairs trading is a popular classic neutral trading strategy in financial market. Deep reinforcement learning (DRL) has been widely used to improve the performance of this strategy. However, most works primarily focused on setting trading signals, but ignored selecting appropriate trading pairs. In this paper, a novel two-level reinforcement learning framework is proposed, where both pair selection and trading thresholds setting are involved. For pair selection, an Extended Option-Critic (EOC) method is utilized, which allows the agent to select trading pair on non-fixed length of time intervals. For trading thresholds setting, a three-agent Multi-Agent Deep Deterministic Policy Gradient (MADDPG) method is used for setting the opening and stop-loss thresholds as well as decide whether to trade. The simulation results in the Chinese futures market demonstrate that our proposed method achieves higher returns compared to traditional methods and popular reinforcement learning approaches.

Simplified expression and recursive algorithm of multi-threshold Tsallis entropy

Image segmentation is an important step in obtaining image information, and it has always been an extremely critical link in the fields of computer vision and pattern recognition. Because of its simplicity and effectiveness, threshold-based image segmentation technology has received significant attention and research by researchers. Tsallis entropy and Renyi entropy are two important global threshold selection methods in image thresholding. They not only have a special correspondence in mathematical expression, but also have a certain relationship in the field of image segmentation. Considering the number of thresholds increases, the complexity of the expression of Tsallis entropy criterion function also increases sharply, which is difficult to apply in multiple thresholds. Given this, this paper proposes a simplified expression of Tsallis entropy with multiple thresholds, which is proved by mathematical induction. Simultaneously, it is concluded that the multiple-threshold segmentation of Tsallis entropy and Renyi entropy with the same parameters is equivalent. To overcome the inconvenience of high calculation cost under multi-threshold and to improve calculation efficiency, recursive algorithms based on simplified Tsallis entropy and Renyi entropy under multi-threshold are given. The calculation costs of the traditional Tsallis entropy and Renyi entropy threshold, simplified Tsallis entropy and Renyi entropy threshold, and the recursive simplified Tsallis entropy and Renyi entropy are compared through experiments. The results have revealed that the calculation time of simplified Tsallis entropy and Renyi entropy is shorter than the traditional methods, and the calculation time of the recursive simplified Tsallis entropy and Renyi entropy is much lower than that of the simplified expression. Finally, we use the three optimization algorithms including particle swarm, differential evolution and Harris hawks algorithm to solve the optimal solutions of the six expressions in the case of multilevel thresholds. Experiments also show that recursive simplified Tsallis entropy and Renyi entropy have the lowest calculation time cost under the same parameters.

Archival skin biopsy specimens as a tool for miRNA-based diagnosis: Technical and post-analytical considerations

Archived specimens, taken by standardized procedures in clinical practice, represent a valuable resource in translational medicine. Their use in retrospective molecular-based studies could provide disease and therapy predictors. Microfluidic array is a user-friendly and cost-effective method allowing profiling of hundreds of microRNAs (miRNAs) from a low amount of RNA. However, even though tissue miRNAs may include potentially robust biomarkers, non-uniformed post-analytical pipelines could hinder translation into clinics. In this study, epidermal RNA from archival skin biopsy specimens was isolated from patients with peripheral neuropathy and healthy individuals. Unbiased miRNA profiling was performed using RT-qPCR-based microfluidic array. We demonstrated that RNA obtained from archival tissue is appropriate for miRNA profiling, providing evidence that different practices in threshold selection could significantly influence the final results. We showed the utility of software-based quality control for amplification curves. We revealed that selection of the most stable reference and the calculation of geometric mean are suitable when utilizing microfluidic arrays without known references. By applying appropriate post-analytical settings, we obtained miRNA profile of human epidermis associated with biological processes and a list of suitable references. Our results, which outline technical and post-analytical considerations, support the broad use of archived specimens for miRNA analysis to unravel disease-specific molecular signatures.

An F-ratio-based method for estimating the number of active sources in MEG.

Magnetoencephalography (MEG) is a powerful technique for studying the human brain function. However, accurately estimating the number of sources that contribute to the MEG recordings remains a challenging problem due to the low signal-to-noise ratio (SNR), the presence of correlated sources, inaccuracies in head modeling, and variations in individual anatomy. To address these issues, our study introduces a robust method for accurately estimating the number of active sources in the brain based on the F-ratio statistical approach, which allows for a comparison between a full model with a higher number of sources and a reduced model with fewer sources. Using this approach, we developed a formal statistical procedure that sequentially increases the number of sources in the multiple dipole localization problem until all sources are found. Our results revealed that the selection of thresholds plays a critical role in determining the method's overall performance, and appropriate thresholds needed to be adjusted for the number of sources and SNR levels, while they remained largely invariant to different inter-source correlations, translational modeling inaccuracies, and different cortical anatomies. By identifying optimal thresholds and validating our F-ratio-based method in simulated, real phantom, and human MEG data, we demonstrated the superiority of our F-ratio-based method over existing state-of-the-art statistical approaches, such as the Akaike Information Criterion (AIC) and Minimum Description Length (MDL). Overall, when tuned for optimal selection of thresholds, our method offers researchers a precise tool to estimate the true number of active brain sources and accurately model brain function.

Supplementation of minerals and vitamins influences optimal targeted selective treatment thresholds for the control of gastro-intestinal nematodes in lambs

Targeted Selective Treatment (TST) is a gastrointestinal nematode (GIN) control strategy where anthelmintic treatment decisions are made at an individual animal level. TST has been proven to reduce anthelmintic use and subsequently slow down anthelmintic resistance development, however questions remain regarding optimal TST methods and their applicability across farms.In this study, the influence of Mineral and Vitamin (MV) supplementation on optimal energy utilisation (EU) TST thresholds was assessed on three Welsh farms. In total, 360 lambs were split into two groups, MV supplemented and control, and were treated with an anthelmintic against GIN at the midway point of the experiment. Lambs that improved their EU efficiency post treatment were deemed to have benefited from anthelmintic treatment. Optimal EU TST thresholds was determined for each treatment group per farm using Youden’s J statistic where the treatment threshold retrospectively exhibiting the greatest combined sensitivity and specificity in correctly identifying lambs benefiting from treatment was deemed to be optimal.Results demonstrated that the optimal EU TST threshold was higher in MV supplemented groups at 0.72, 0.71 and 0.56 versus 0.58, 0.67, 0.51 for control groups on each respective farm. Identification of lambs for TST was more effective when using an optimised EU TST threshold, compared to when using the standard EU TST threshold of 0.66.The study highlights that applying standard EU TST thresholds may not be appropriate on all commercial farms with factors including MV status as noted in this study likely to influence optimal EU TST thresholds. Additional refinement of TST systems can further strengthen their applicability across sheep flocks.

The impact of teeth and dental restorations on gray value distribution in cone-beam computer tomography: a pilot study

PurposeTo investigate the influence of teeth and dental restorations on the facial skeleton's gray value distributions in cone-beam computed tomography (CBCT).MethodsGray value selection for the upper and lower jaw segmentation was performed in 40 patients. In total, CBCT data of 20 maxillae and 20 mandibles, ten partial edentulous and ten fully edentulous in each jaw, respectively, were evaluated using two different gray value selection procedures: manual lower threshold selection and automated lower threshold selection. Two sample t tests, linear regression models, linear mixed models, and Pearson's correlation coefficients were computed to evaluate the influence of teeth, dental restorations, and threshold selection procedures on gray value distributions.ResultsManual threshold selection resulted in significantly different gray values in the fully and partially edentulous mandible. (p = 0.015, difference 123). In automated threshold selection, only tendencies to different gray values in fully edentulous compared to partially edentulous jaws were observed (difference: 58–75). Significantly different gray values were evaluated for threshold selection approaches, independent of the dental situation of the analyzed jaw. No significant correlation between the number of teeth and gray values was assessed, but a trend towards higher gray values in patients with more teeth was noted.ConclusionsStandard gray values derived from CT imaging do not apply for threshold-based bone segmentation in CBCT. Teeth influence gray values and segmentation results. Inaccurate bone segmentation may result in ill-fitting surgical guides produced on CBCT data and misinterpreting bone density, which is crucial for selecting surgical protocols.Graphical Created with BioRender.com

Assessment of Signal Similarity Index Based Line Protection Scheme for Systems With Inverter Based Resources

This article evaluates the performance of a signal similarity index (SSI) based relaying scheme for fault detection in lines connected to inverter-based resources (IBRs). During fault events, the IBRs respond differently than conventional synchronous machine based sources. Due to various control schemes and power electronic protection implemented for IBRs, fault current is limited leading to failure in fault detection and isolation by traditional protection methods. In this article, six different SSIs are calculated using time-domain current signals at both the ends of the protected line and the performance of SSI-based line protection philosophy is evaluated for various operating conditions. The relaying schemes are extensively tested on two different test systems. The test results are utilized to perform a comparative assessment of the SSI-based protection scheme. The tradeoff between the performance of the relaying scheme based on processing time, fault resistance, threshold selection, presence of noise, and presence of missing data are demonstrated thoroughly on three different test systems.

Robust Audio Copy-Move Forgery Detection Using Constant Q Spectral Sketches and GA-SVM

Audio recordings used as evidence have become increasingly important to litigation. Before their admissibility as evidence, an audio forensic expert is often required to help determine whether the submitted audio recordings are altered or authentic. Within this field, the copy-move forgery detection (CMFD), which focuses on finding possible forgeries that are derived from the same audio recording, has been an urgent problem in blind audio forensics. However, most of the existing methods require idealistic pre-segmentation and artificial threshold selection to calculate the similarity between segments, which may result in serious misleading and misjudgment especially on high frequency words. In this work, we present a robust method for detecting and locating an audio copy-move forgery on the basis of constant Q spectral sketches (CQSS) and the integration of a customised genetic algorithm (GA) and support vector machine (SVM). Specifically, the CQSS features are first extracted by averaging the logarithm of the squared-magnitude constant Q transform. Then, the CQSS feature set is automatically optimised by a customised GA combined with SVM to obtain the best feature subset and classification model at the same time. Finally, the integrated method, named CQSS-GA-SVM, is evaluated against the state-of-the-art approaches to blind detection of copy-move forgeries on real-world copy-move datasets with read English and Chinese corpus, respectively. The experimental results demonstrate that the proposed CQSS-GA-SVM exhibits significantly high robustness against post-processing based anti-forensics attacks and adaptability to the changes of the duplicated segment duration, the training set size, the recording length, and the forgery type, which may be beneficial to improving the work efficiency of audio forensic experts.

Automated, Reliable Zero-Day Malware Detection Based on Autoencoding Architecture

While a body of studies has been carried out for malware detection with its significance, they are often limited to known malware patterns due to the reliance on signature-based or supervised learning approaches. The semi-supervised learning approach would be an option for identifying previously unseen patterns (i.e., zero-day detection); however, our preliminary study reveals critical limitations from existing methods, including (i) the profiling-based approach using an autoencoder can provide better detection but is sensitive to the threshold setting, and (ii) one-class (OC) classification does not require a manual threshold discovery but may be limited with low detection rates. In this paper, we present a new detection method incorporating the concept of autoencoding and OC classification, designed to benefit from strong abstraction by neural networks (using an autoencoder) and the removal of the complex threshold selection (using an OC classifier). For this combined architecture, a challenge is concurrent training of the autoencoder and the OC classifier, which may cause an ill-suited learner due to no reference to malware instances. To this end, we introduce a new model selection method that discovers well-optimized models from a variety of combinations. The experimental results performed with public malware datasets (Meraz’18 and Drebin) show the effectiveness of our presented methods with up to 97.1% accuracy, comparable to the supervised learning-based detection. We also examine the impact of evading attacks using adversarial attack tools, the result of which shows resilience to malware variants with over 99% detection rates.

Design of some Apigenin derivatives as selective DPP-IV Inhibitors by Pharmacophore Modelling and its Validation through Molecular Dynamics Simulation

One of the most common forms of diabetes, type-2 diabetes mellitus is a chronic metabolic condition brought on by problems with insulin production or insulin resistance. By modifying blood glucose homeostasis, dipeptidyl peptidase-IV (DPP-IV) inhibition has the potential to treat type 2 diabetes. The quick development of computational drug design provided a fantastic opportunity to both discover and forecast the effectiveness of potential DPP-IV inhibitors. The current investigation supports and verifies the identification of new Apigenin derivatives as selective DPP-IV inhibitors. Schrodinger software was used to perform virtual screening (VS) of created compounds against DPP-IV, and the most promising hits were chosen. Out of 110 substances, 56 new apigenin derivatives were chosen for additional docking investigations using Glide based on their selectivity threshold followed by validation by pose selection method which is found to be below 2 Ǻ based on their superimposion with co-crystallized lgand. The DPP-IV protein was used as the target of molecular dynamic (MD) simulation studies to evaluate the correct binding mechanisms and stability of their complexes with enzyme. Structural chemistry was utilised to guide and considerably speed up the drug discovery process during the back-up development stages. It was also used to analyse in-silico suggestions and screening results, as well as to come up with fresh ideas. The study thus demonstrates the possibility of apigenin derivatives as highly specific DPP-IV inhibitors.

YOLO Network with a Circular Bounding Box to Classify the Flowering Degree of Chrysanthemum

Detecting objects in digital images is challenging in computer vision, traditionally requiring manual threshold selection. However, object detection has improved significantly with convolutional neural networks (CNNs), and other advanced algorithms, like region-based convolutional neural networks (R-CNNs) and you only look once (YOLO). Deep learning methods have various applications in agriculture, including detecting pests, diseases, and fruit quality. We propose a lightweight YOLOv4-Tiny-based object detection system with a circular bounding box to accurately determine chrysanthemum flower harvest time. The proposed network in this study uses a circular bounding box to accurately classify the degree of chrysanthemums blooming and detect circular objects effectively, showing better results than the network with the traditional rectangular bounding box. The proposed network has excellent scalability and can be applied to recognize general objects in a circular form.

A Practical Guide for Left Adrenal Vein Sampling in Patients with Left Renal Vein Variants.

The presence of left renal vein (LRV) variants can increase the complexity of adrenal vein sampling (AVS), an already technically demanding procedure. While AVS literature often focuses on the right adrenal vein, an understanding of common LRV variants, their relationship with the left adrenal vein, and principles for successful catheterization can facilitate AVS. This guide provides practical, technical tips for AVS for duplicated (Du), circumaortic (Ca), and retroaortic (Ra) LRVs. AVS cases were identified at a single institution (June 2009-March 2023) based on adrenophrenic trunk drainage relative to variant LRVs. Available cross-sectional imaging was reviewed to evaluate LRV anatomy pre-procedure. Twenty-seven cases (1 DuLRV, 13 CaLRVs, and 13 RaLRVs) were identified. Diagnostic AVS was confirmed by a threshold selectivity index. Literature on LRV anatomic variants was also reviewed. Based on the authors' experience and literature review, the following principles can guide AVS in the setting of LRV variants. In the presence of DuLRV or CaLRV, the left adrenal vein invariably drains into a normally positioned, pre-aortic LRV limb, so AVS can proceed as expected with a Simmons as the catheter of choice. In contrast, a LAV draining into a RaLRV may require a hockey stick-like catheter, or in rare cases a microcatheter, for selecting and sampling, due to the longer RaLRV course, which usually drains into the IVC more inferiorly and can be stenotic where the aorta crosses. Knowing the presence and understanding the anatomy of LRV variants can facilitate an efficient AVS.

Optimal threshold selection for segmentation of Chest X-Ray images using opposition-based swarm-inspired algorithm for diagnosis of pneumonia

Optimal threshold selection for segmentation of Chest X-Ray images using opposition-based swarm-inspired algorithm for diagnosis of pneumonia

Nonstationary frequency analysis and uncertainty quantification for extreme low lake levels in a large river-lake-catchment system

Extreme hydrological events have become increasingly frequent on a global scale. The middle Yangtze River also faces a substantial challenge in dealing with extreme flooding and drought. However, the long-term characteristics of the extreme hydrological regime have not yet been adequately recognized. Moreover, there is uncertainty in the extreme value estimation, and this uncertainty needs to be distinguished and quantified. In this study, we investigated the nonstationary frequency characteristics of extreme low lake levels (ELLLs), taking the Poyang Lake as an example. Daily lake levels from 1960 to 2022 were utilized to estimate the return level using the generalized Pareto distribution (GPD). The uncertainty from three sources, i.e., the parameter estimator, threshold selection, and covariate, was quantified via variance decomposition. The results indicate that (1) the parameter estimator is the predominant source of uncertainty, with a contribution rate of approximately 87 %. The total uncertainty of the covariate, threshold, and interaction term is only 13 %. (2) Two indexes, namely the annual minimum water level (WLmin) and the days with peak over the 90 % threshold per year (DPOT90), decreased (0.01–0.03 m/year) and increased (0.17–1.39 days/year), respectively, indicating a progressively severe drought trend for Poyang Lake. (3) The return level with return period of 5 to 100 years significantly decreased after the early 21st century. A large spatial heterogeneity was identified for the variation in the return level, and the change rate of the return level with a 100-year return period ranged from 5 % to 40 % for the whole lake. (4) The ELLLs had a stronger correlation with the catchment discharge than with the Yangtze River discharge and the large-scale atmospheric circulation indices. This study provides a methodology with reduced uncertainty for nonstationary frequency analysis (NFA) of ELLLs exemplified in large river–lake systems.

PSF toolkit: an R package for pathway curation and topology-aware analysis.

Most high throughput genomic data analysis pipelines currently rely on over-representation or gene set enrichment analysis (ORA/GSEA) approaches for functional analysis. In contrast, topology-based pathway analysis methods, which offer a more biologically informed perspective by incorporating interaction and topology information, have remained underutilized and inaccessible due to various limiting factors. These methods heavily rely on the quality of pathway topologies and often utilize predefined topologies from databases without assessing their correctness. To address these issues and make topology-aware pathway analysis more accessible and flexible, we introduce the PSF (Pathway Signal Flow) toolkit R package. Our toolkit integrates pathway curation and topology-based analysis, providing interactive and command-line tools that facilitate pathway importation, correction, and modification from diverse sources. This enables users to perform topology-based pathway signal flow analysis in both interactive and command-line modes. To showcase the toolkit's usability, we curated 36 KEGG signaling pathways and conducted several use-case studies, comparing our method with ORA and the topology-based signaling pathway impact analysis (SPIA) method. The results demonstrate that the algorithm can effectively identify ORA enriched pathways while providing more detailed branch-level information. Moreover, in contrast to the SPIA method, it offers the advantage of being cut-off free and less susceptible to the variability caused by selection thresholds. By combining pathway curation and topology-based analysis, the PSF toolkit enhances the quality, flexibility, and accessibility of topology-aware pathway analysis. Researchers can now easily import pathways from various sources, correct and modify them as needed, and perform detailed topology-based pathway signal flow analysis. In summary, our PSF toolkit offers an integrated solution that addresses the limitations of current topology-based pathway analysis methods. By providing interactive and command-line tools for pathway curation and topology-based analysis, we empower researchers to conduct comprehensive pathway analyses across a wide range of applications.