Low-quality Samples Research Articles

Image Recognition Method for Corn Northern Leaf Blight of Unmanned Aerial Vehicles Based on Deep Learning

Aiming at the problems of high complexity, false detection and missing detection of maize big spot based on UAV image, an improved algorithm for maize big spot detection was proposed. The algorithm is based on EMA and the improved YOLOv8-BiFPN feature pyramid network. By adding the efficient multi-scale EMA attention mechanism module, the capturing ability of detail information is improved, thus enhancing the feature extraction ability of the model. The YOLOv8 structure is improved into YOLOV8-BIFPN feature pyramid network by integrating BiFPN structure, which can extract context information more efficiently. By introducing WloU loss function, low quality samples in training data are filtered effectively, and the generalization ability of the model is improved. In this paper, the accuracy P was increased by 2.7%, the recall rate R was increased by 4.7%, the average accuracy mAP50 was increased by 3.1%, and the model size was reduced by 2.78M. Compared with other YOLO algorithms, the proposed method has significant advantages in the detection of corn big spot disease.

DVCW-YOLO for Printed Circuit Board Surface Defect Detection

The accurate and efficient detection of printed circuit board (PCB) surface defects is crucial to the electronic information manufacturing industry. However, current approaches to PCB defect detection face challenges, including large model sizes and difficulties in balancing detection accuracy with speed. To address these challenges, this paper proposes a novel PCB surface defect detection algorithm, named DVCW-YOLO. First, all standard convolutions in the backbone and neck networks of YOLOv8n are replaced with lightweight DWConv convolutions. In addition, a self-designed C2fCBAM module is introduced to the backbone network for extracting features. Next, within the neck structure, the C2f module is substituted with the more lightweight VOVGSCSP module, thereby reducing model redundancy, simplifying model complexity, and enhancing detection speed. By enhancing prominent features and suppressing less important ones, this modification allows the model to better focus on key regions, thereby improving feature representation capabilities. Finally, the WIoU loss function is implemented to replace the traditional CIoU function in YOLOv8n. This adjustment addresses issues related to low generalization and poor detection performance for small objects or complex backgrounds, while also mitigating the impact of low-quality or extreme samples on model accuracy. Experimental results demonstrate that the DVCW-YOLO model achieves a mean average precision (mAP) of 99.3% and a detection speed of 43.3 frames per second (FPS), which represent improvements of 4% and 4.08%, respectively, over the YOLOv8n model. These results confirm that the proposed model meets the real-time PCB defect detection requirements of small and medium-sized enterprises.

Inferring Distant Relationships From Dense SNP Data Utilizing Two Genealogy Algorithms.

A highly esteemed method known as investigative genetic genealogy (IGG) has been developed to identify DNA samples from forensic crime scenes and human remains of disaster victims. With the advent of next-generation sequencing, it is now feasible to access information on millions of SNPs typed in a single sequencing run that fulfill the requirements for kinship inference. However, challenges such as the poor quality of forensic samples, the high cost associated with sequencing technology, and privacy concerns regarding large-scale genetic databases remain unresolved in this field. In the present study, we validated the identification of relationships up to the seventh degree using two genealogy algorithms (IBIS and KING) under various parameter settings. This was accomplished through whole genome sequencing data derived from two southern Chinese Han pedigrees during an initial phase, while also exploring workflows adapted for low-quality samples. To achieve this objective, low-coverage whole genome sequencing data were downsampled from high-coverage original datasets; additionally, mimic SNP array data-containing less information but offering greater accessibility-were prepared as reference samples. Through a series of experimental analyses, we not only validate the applicability of selected processing procedures and inference tools for low-coverage samples but also proposed that a meticulously crafted site filtering strategy can significantly improve the accuracy of kinship identification. This acknowledges the necessity for further systematic evidence in future research endeavors.

Improved Algorithm for Vehicle Bottom Safety Detection Based on YOLOv8n: PSP-YOLO

The safety inspection system for the underside of vehicles demands both high detection speed and accuracy, necessitating a model with small parameters and high network efficiency. To address these issues and improve the real-time performance of under-vehicle safety inspection, the YOLOv8n object-detection algorithm was enhanced, resulting in the PSP-YOLO algorithm. Firstly, the improved PSP-YOLO network introduces a P2 small-object detection head, which is particularly effective in detecting tiny objects, such as small dangerous objects under vehicles, thereby significantly enhancing the ability to detect such small targets. Secondly, the Space-to-Depth Conv (SPD-Conv) module was introduced into the backbone network, which improves the detection performance of low-quality samples and small objects while ensuring high precision and accurate localization of small targets in images. Finally, the Partial Convolution (PConv) lightweight module was also added to the feature fusion network, effectively reducing the model’s parameters and computational load, conserving computing resources, and improving detection speed and accuracy. Experimental results demonstrate that the modified model increased mAP@0.5 by 3.9% and mAP@0.5-0.95 by 3.6% on the dataset used, reduced the number of parameters to 2.53 M, and achieved a detection speed of 104.6 f/s. Therefore, the improved YOLOv8n algorithm significantly enhances the detection performance of dangerous objects, meeting the high-precision and real-time requirements for vehicle underside safety inspections.

A Pipeline and Recommendations for Population and Individual Diagnostic SNP Selection in Non-Model Species.

Despite substantial reductions in the cost of sequencing over the last decade, genetic panels remain relevant due to their cost-effectiveness and flexibility across a variety of sample types. In particular, single nucleotide polymorphism (SNP) panels are increasingly favoured for conservation applications. SNP panels are often used because of their adaptability, effectiveness with low-quality samples, and cost-efficiency for population monitoring and forensics. However, the selection of diagnostic SNPs for population assignment and individual identification can be challenging. The consequences of poor SNP selection are under-powered panels, inaccurate results, and monetary loss. Here, we develop a novel anduser-friendly SNP selection pipeline (mPCRselect) that can be used to select SNPs for population assignment and/or individual identification. mPCRselect allows any researcher, who has sufficient SNP-level data, to design a successful and cost-effective SNP panel for a diploid species of conservation concern.

Electroacupuncture versus 5-HT4 receptor agonist for functional constipation: A systematic review and meta-analysis of randomized controlled trials.

Functional constipation (FC) has been found as a chronic gastrointestinal disease that is commonly diagnosed in patients. However, patients have a low satisfaction level with the treatment of constipation drugs (e.g., 5-HT4 agonists). A meta-analysis was performed to compare the efficacy and safety between electroacupuncture and 5-HT4 agonists. The included study were randomized controlled trials (RCTs), in which EA was used in the experimental group and 5-HT4 receptor agonist was used in the control group. Four English databases (PubMed, Cochrane Library, Web of Science, Embase) and 4 Chinese databases (China National Knowledge Infrastructure, CBM, WanFang, VIP) were searched. Relevant studies retrieved were published before September 30, 2024. The risk of bias was assessed by tool of Cochrane and GRADEpro. The Review Manager 5.4 was used for analyzing Data analysis, and Endnote X9 for screening studies. In this paper, we included 12 studies, involving 1473 participants. We found that EA significantly improved patient assessment of cab quality of life questionnaire (PAC-QOL) (MD = -0.52, P = .03), self-rating anxiety scale (SAS) (MD = -3.00, P < .00001) and self-rating depression scale (SDS) (MD = -4.13, P < .00001) compared with 5-HT4 receptor agonists. In addition, we failed to identify any significant difference in Stool consistency, the number of weekly complete spontaneous bowel movements and weekly spontaneous bowel movements (SBMs) between the 2 groups. EA has been indicated to be better than 5-HT4 receptor agonists since it can more effectively improve FC patients' life quality and mental state without an increased risk of adverse even. However, the previous evidence is characterized by low quality and small sample size, which should be further confirmed by high-quality and large-sample multicenter RCTs.

Abstract 4117767: Interplay Between ST2 Gene Downregulation and Inflammatory Processes in Hypertrophic Cardiomyopathy

Introduction: Hypertrophic Cardiomyopathy (HCM) is an inherited heart disease and the pathogenesis of HCM involves genetic mutations, hemodynamic stress, and metabolic factors, with myocardial fibrosis playing a crucial role in severe clinical events. IL-33/ST2 signaling pathway known for its roles in immune response and tissue repair, participates in cardiac protection and anti-cardiac fibrosis in heart failure. The role of ST2 in HCM remains unclear, and IL-33/ST2 pathway and broader inflammatory responses may be critical in HCM. Hypothesis: ST2 gene downregulation and dysregulation of inflammatory pathways represent a novel mechanism in HCM. Methods: We re-analyzed the raw whole transcriptome sequencing data using a comprehensive bioinformatics pipeline from 9 GEO datasets. After removing low-quality samples, myocardial tissue samples from 109 HCM patients and 210 non-HCM controls were included. HTSeq-count was used to obtain the read count of each gene. Differential expression analyses were conducted using DESeq2. Spearman correlation analysis was performed for ST2 with all other protein-coding genes. Gene set enrichment analysis (GSEA) was performed to explore the biological significance of ST2 related genes. CIBERSORTx was used to estimate the immune cell composition of heart tissues. Additionally, network analyses and statistical correlations were conducted. Results: Our analysis identified significant downregulation of the ST2 gene in HCM patients compared to controls (log2FC = -5.0, p = 2.7E-147). In total, 5,180 upregulated and 741 downregulated genes were detected. Additionally, 208 genes were significantly correlated with the expression of ST2. GSEA based on the correlation coefficients revealed a significant increase in immune response, inflammation, and IFN-γ pathway activity, along with a decrease in the expression of genes related to sarcomeric function, cardiac morphogenesis, and metabolism in HCM. Notably, the significant inverse correlation between ST2 expression and regulatory T cells (r = -0.34) and a positive correlation with neutrophils (r = 0.39) suggest that ST2 may modulate immune cell populations in HCM. Pathway analysis indicated ST2's central role in networks involving inflammatory and fibrotic responses. Conclusion: The interplay between ST2 and inflammatory pathways may drive myocardial remodeling and fibrosis in HCM. This study paves a new way for investigating pathogenic mechanisms and therapeutic strategies for HCM.

A Pseudo-Labeling Multi-Screening-Based Semi-Supervised Learning Method for Few-Shot Fault Diagnosis.

In few-shot fault diagnosis tasks in which the effective label samples are scarce, the existing semi-supervised learning (SSL)-based methods have obtained impressive results. However, in industry, some low-quality label samples are hidden in the collected dataset, which can cause a serious shift in model training and lead to the performance of SSL-based method degradation. To address this issue, the latest prototypical network-based SSL techniques are studied. However, most prototypical network-based scenarios consider that each sample has the same contribution to the class prototype, which ignores the impact of individual differences. This article proposes a new SSL method based on pseudo-labeling multi-screening for few-shot bearing fault diagnosis. In the proposed work, a pseudo-labeling multi-screening strategy is explored to accurately screen the pseudo-labeling for improving the generalization ability of the prototypical network. In addition, the AdaBoost adaptation-based weighted technique is employed to obtain accurate class prototypes by clustering multiple samples, improving the performance that deteriorated by low-quality samples. Specifically, the squeeze and excitation block technique is used to enhance the useful feature information and suppress non-useful feature information for extracting accuracy features. Finally, three well-known bearing datasets are selected to verify the effectiveness of the proposed method. The experiments illustrated that our method can receive better performance than that of the state-of-the-art methods.

The Research of Intra-Pulse Modulated Signal Recognition of Radar Emitter under Few-Shot Learning Condition Based on Multimodal Fusion

Radar radiation source recognition is critical for the reliable operation of radar communication systems. However, in increasingly complex electromagnetic environments, traditional identification methods face significant limitations. These methods often struggle with high noise levels and diverse modulation types, making it difficult to maintain accuracy, especially when the Signal-to-Noise Ratio (SNR) is low or the available training data are limited. These difficulties are further intensified by the necessity to generalize in environments characterized by a substantial quantity of noisy, low-quality signal samples while being constrained by a limited number of desirable high-quality training samples. To more effectively address these issues, this paper proposes a novel approach utilizing Model-Agnostic Meta-Learning (MAML) to enhance model adaptability in few-shot learning scenarios, allowing the model to quickly learn with limited data and optimize parameters effectively. Furthermore, a multimodal fusion neural network, DCFANet, is designed, incorporating residual blocks, squeeze and excitation blocks, and a multi-scale CNN, to fuse I/Q waveform data and time–frequency image data for more comprehensive feature extraction. Our model enables more robust signal recognition, even when the signal quality is severely degraded by noise or when only a few examples of a signal type are available. Testing on 13 intra-pulse modulated signals in an Additive White Gaussian Noise (AWGN) environment across SNRs ranging from −20 to 10 dB demonstrated the approach’s effectiveness. Particularly, under a 5−way5−shot setting, the model achieves high classification accuracy even at −10dB SNR. Our research underscores the model’s ability to address the key challenges of radar emitter signal recognition in low-SNR and data-scarce conditions, demonstrating its strong adaptability and effectiveness in complex, real-world electromagnetic environments.

Active Learning for Deep Visual Tracking.

Convolutional neural networks (CNNs) have been successfully applied to the single target tracking task in recent years. Generally, training a deep CNN model requires numerous labeled training samples, and the number and quality of these samples directly affect the representational capability of the trained model. However, this approach is restrictive in practice, because manually labeling such a large number of training samples is time-consuming and prohibitively expensive. In this article, we propose an active learning method for deep visual tracking, which selects and annotates the unlabeled samples to train the deep CNN model. Under the guidance of active learning, the tracker based on the trained deep CNN model can achieve competitive tracking performance while reducing the labeling cost. More specifically, to ensure the diversity of selected samples, we propose an active learning method based on multiframe collaboration to select those training samples that should be and need to be annotated. Meanwhile, considering the representativeness of these selected samples, we adopt a nearest-neighbor discrimination method based on the average nearest-neighbor distance to screen isolated samples and low-quality samples. Therefore, the training samples' subset selected based on our method requires only a given budget to maintain the diversity and representativeness of the entire sample set. Furthermore, we adopt a Tversky loss to improve the bounding box estimation of our tracker, which can ensure that the tracker achieves more accurate target states. Extensive experimental results confirm that our active-learning-based tracker (ALT) achieves competitive tracking accuracy and speed compared with state-of-the-art trackers on the seven most challenging evaluation benchmarks. Project website: https://sites.google.com/view/altrack/.

Landsat image classification using a deep learning model and multiple-source training samples

ABSTRACT Public release of the entire Landsat data archive and cloud-based geocomputation platforms have greatly facilitated land-cover mapping worldwide. The acquisition of training samples presents a significant challenge in mapping. In this study, we have developed an approach based on transfer learning to creatively use existing global land-cover products and training samples to resolve the training sample issue. We pretrained a deep neural network (DNN) model using low-quality samples automatically generated from existing global land-cover products. The pretrained model was then fine-tuned using high-quality training samples gathered from the training samples shared by authors of existing studies. We implemented this approach to generate a land-cover map for Beijing City, China, in 2015. Using the fine-tuned DNN model, we achieved an overall accuracy (OA) of 86.4% and a kappa coefficient of 0.796, based on independent validation samples. The accuracy was the same and even surpassed that of existing land-cover maps for Beijing. The fine-tuned DNN outperformed the random forest (RF) (OA = 70.2%, kappa = 0.588) and support vector machine (SVM) models (OA = 68.7%, kappa = 0.555) using the same training samples, showing that the need for training samples in land-cover classification can be met by combining deep learning models, existing global land-cover products, and recycled training samples.

YOLO-EF: A lightweight YOLOv7 PCB defect detection algorithm

Abstract This study addresses the challenge of accurate PCB defect detection, considering the computational complexity and excessive model parameters associated with existing methods. To mitigate these issues, we propose an enhanced PCB surface defect detection method called YOLO-EF, which is built upon the YOLOv7 architecture. Firstly, we replace the backbone of YOLOv7 with the lightweight Efficient FormerV2 module. This replacement reduces both the model volume and computational requirements. Secondly, the dynamic head module is introduced to unify the target detection head and improve its performance. Finally, the Wise-IoUv3 function is used to suppress the influence of low-quality samples on the model performance. The experimental results show that the proposed model improves by 2.85% in mAP over the original model. Model computation is reduced from 106.5 GFLOPs to 44.3 GFLOPs. Our YOLO-EF model achieves an optimal balance in terms of precision, computational efficiency, and detection speed.

Shotgun DNA sequencing for human identification: Dynamic SNP selection and likelihood ratio calculations accounting for errors

Shotgun sequencing is a DNA analysis method that potentially determines the nucleotide sequence of every DNA fragment in a sample, unlike PCR-based genotyping methods that is widely used in forensic genetics and targets predefined short tandem repeats (STRs) or predefined single nucleotide polymorphisms (SNPs). Shotgun DNA sequencing is particularly useful for highly degraded low-quality DNA samples, such as ancient samples or those from crime scenes. Here, we developed a statistical model for human identification using shotgun sequencing data and developed formulas for calculating the evidential weight as a likelihood ratio (LR). The model uses a dynamic set of binary SNP loci and takes the error rate from shotgun sequencing into consideration in a probabilistic manner. To our knowledge, the method is the first to make this possible. Results from replicated shotgun sequencing of buccal swabs (high-quality samples) and hair samples (low-quality samples) were arranged in a genotype-call confusion matrix to estimate the calling error probability by maximum likelihood and Bayesian inference. Different genotype quality filters may be applied to account for genotyping errors. An error probability of zero resulted in the commonly used LR formula for the weight of evidence. Error probabilities above zero reduced the LR contribution of matching genotypes and increased the LR in the case of a mismatch between the genotypes of the trace and the person of interest. In the latter scenario, the LR increased from zero (occurring when the error probability was zero) to low positive values, which allow for the possibility that the mismatch may be due to genotyping errors. We developed an open-source R package, wgsLR, which implements the method, including estimation of the calling error probability and calculation of LR values. The R package includes all formulas used in this paper and the functionalities to generate the formulas.

Research on gear flank surface defect detection of automotive transmissions based on improved YOLOv8

AbstractIn industrial production, the manufacturing processes may introduce defects on the gear flanks of transmission gears, potentially leading to premature failures and diminished performance. The early detection and precise assessment of surface defects on transmission gear flanks are critical for maintaining the safety, reliability, and cost-effectiveness of automobiles. At present, the principal approach for identifying defects on automotive transmission gear flanks predominantly involves manual visual inspections, supplemented by fluorescent magnetic particle testing. However, this approach suffers from low accuracy and efficiency. Consequently, this paper presents a defect detection algorithm that leverages an enhanced YOLOv8 model to facilitate the efficient detection of surface defects on automotive transmission gear flanks. Initially, the collected image data underwent data augmentation and exploratory analysis, which informed targeted enhancements. Subsequently, the YOLOv8 algorithm was thoroughly examined. The spatial pyramid pooling layer efficient architecture was incorporated into the backbone network, and the Deformable Convolutional Networks v4 module was integrated to boost the model’s capability in detecting irregular defects. In the neck network, the BiFormer attention mechanism was implemented to enhance detection performance for small-scale defects. Moreover, the newly developed modified adaptive structure feature fusion MASFF_Head structure was adopted as the detection head to augment detection efficacy for multi-scale defects. Additionally, the bounding box loss function was substituted with the Wise-Intersection over Union (WIoU) loss function to improve performance on low-quality samples. Experimental results demonstrated that the mean average precision (mAP@0.5) of the refined YOLOv8 network model reached 86.1%, marking a 2.8% increase over the original model and significantly boosting detection accuracy. When compared to other deep learning models, the enhanced YOLOv8 model exhibits considerable superiority in terms of detection precision and efficiency. The precision (P) value and recall (R) value achieved were 82.9% and 80.8%, respectively, with a detection time of 21.6 milliseconds. This underscores the method’s effectiveness and reliability in detecting automotive transmission gear defects, underscoring its pivotal role in facilitating automated detection processes on industrial production lines.

Lightweight Single-Stage Ship Object Detection Algorithm for Unmanned Surface Vessels Based on Improved YOLOv5.

Object detection is applied extensively in various domains, including industrial manufacturing, road traffic management, warehousing and logistics, and healthcare. In ship object detection tasks, detection networks are frequently deployed on devices with limited computational resources, e.g., unmanned surface vessels. This creates a need to balance accuracy with a low parameter count and low computational load. This paper proposes an improved object detection network based on YOLOv5. To reduce the model parameter count and computational load, we utilize an enhanced ShuffleNetV2 network as the backbone. In addition, a split-DLKA module is devised and implemented in the small object detection layer to improve detection accuracy. Finally, we introduce the WIOUv3 loss function to minimize the impact of low-quality samples on the model. Experiments conducted on the SeaShips dataset demonstrate that the proposed method reduces parameters by 71% and computational load by 58% compared to YOLOv5s. In addition, the proposed method increases the mAP@0.5 and mAP@0.5:0.95 values by 3.9% and 3.3%, respectively. Thus, the proposed method exhibits excellent performance in both real-time processing and accuracy.

Applications and Performance of Precision ID GlobalFiler NGS STR, Identity, and Ancestry Panels in Forensic Genetics.

Short Tandem Repeat (STR) testing via capillary electrophoresis is undoubtedly the most popular forensic genetic testing method. However, its low multiplexing capabilities and limited performance with challenging samples are among the factors pushing scientists towards new technologies. Next-generation sequencing (NGS) methods overcome some of these limitations while also enabling the testing of Single-Nucleotide Polymorphisms (SNPs). Nonetheless, these methods are still under optimization, and their adoption into practice is limited. Among the available kits, Thermo Fisher Scientific (Waltham, MA, USA) produces three Precision ID Panels: GlobalFiler NGS STR, Identity, and Ancestry. A clear review of these kits, providing information useful for the promotion of their use, is, however, lacking. To close the gap, a literature review was performed to investigate the popularity, applications, and performance of these kits. Following the PRISMA guidelines, 89 publications produced since 2015 were identified. China was the most active country in the field, and the Identity Panel was the most researched. All kits appeared robust and useful for low-quality and low-quantity samples, while performance with mixtures varied. The need for more population data was highlighted, as well as further research surrounding variables affecting the quality of the sequencing results.

Overlooked poor-quality patient samples in sequencing data impair reproducibility of published clinically relevant datasets

BackgroundReproducibility is a major concern in biomedical studies, and existing publication guidelines do not solve the problem. Batch effects and quality imbalances between groups of biological samples are major factors hampering reproducibility. Yet, the latter is rarely considered in the scientific literature.ResultsOur analysis uses 40 clinically relevant RNA-seq datasets to quantify the impact of quality imbalance between groups of samples on the reproducibility of gene expression studies. High-quality imbalance is frequent (14 datasets; 35%), and hundreds of quality markers are present in more than 50% of the datasets. Enrichment analysis suggests common stress-driven effects among the low-quality samples and highlights a complementary role of transcription factors and miRNAs to regulate stress response. Preliminary ChIP-seq results show similar trends. Quality imbalance has an impact on the number of differential genes derived by comparing control to disease samples (the higher the imbalance, the higher the number of genes), on the proportion of quality markers in top differential genes (the higher the imbalance, the higher the proportion; up to 22%) and on the proportion of known disease genes in top differential genes (the higher the imbalance, the lower the proportion). We show that removing outliers based on their quality score improves the resulting downstream analysis.ConclusionsThanks to a stringent selection of well-designed datasets, we demonstrate that quality imbalance between groups of samples can significantly reduce the relevance of differential genes, consequently reducing reproducibility between studies. Appropriate experimental design and analysis methods can substantially reduce the problem.

Mechanisms and elemental partitioning during simultaneous dephosphorization and reduction of Fe-O-P melts by hydrogen plasma

The major obstacle to render steelmaking more sustainable is undoubtedly the decarbonization of its process chains. Accompanied by this challenge, the scarcity of high-grade iron ores to be exploited as feedstock is a near reality. This creates a massive dilemma that will force steelmakers to produce green steel from low-grade iron ores. The hydrogen plasma smelting reduction of iron ores (HPSR) emerges as an attractive CO2-lean pathway to produce iron, where the ore is exposed to a reducing hydrogen-containing plasma (Ar-10 % H2) to get simultaneously melted and reduced. Here, we investigate the reduction of low-quality and low-cost model hematite ore samples containing 0.79 wt.% P via HPSR using an electric arc furnace (EAF). The dephosphorization mechanism of the ore was monitored by inspecting the evolving microstructure of the fast-solidified samples, condensed gas on the surface of electrostatic filters installed inside the EAF, as well as quantifying the corresponding P concentration in all constituents (i.e., unreduced oxides, iron and gas). We found that most of the P evaporates already in the beginning of the process (2 min) where the global O concentration in the melt varies from ∼29 to 21 wt.%. The remaining quantities of P, together with other minor gangue-related impurities (especially Si), allocates within the interdendritic domains of the partially reduced samples, as revealed by high-resolution scanning electron microscopy (SEM), energy dispersive X-ray analysis (EDX) and atom probe tomography (APT). Complementary thermodynamic modelling underpins the experimental observations, revealing that the evaporation of P is a thermodynamically-driven process, and it is dependent on the oxygen content of the melt. About 97 % dephosphorization was achieved, and only residual amounts of P were found within metallic iron domains, as revealed by APT and chemical analysis of the bulk iron (0.07 wt.% P). Our work aims at providing novel scientific-based perspectives in iron and steelmaking, permitting the production of clean and sustainable iron without the need for secondary metallurgical steps to remove undesired impurities.

An improved you only look once algorithm for pronuclei and blastomeres localization

Pronuclei and blastomeres are key structures in early embryonic development, and by localizing these structures simultaneously, the developmental state of the embryo can be assessed more comprehensively. However, there are several unavoidable problems in localization tasks due to the biological characteristics of embryos, including blastomeres overlap, pronuclei overlap, and similarity between pronuclei and background. In this study, we propose a novel localization network for pronuclei and blastomeres, which can solve these localization problems. Firstly, to address the issues related to the overlap of pronuclei and blastomeres, as well as the pronuclei and background similarity problem, we put the vision transformer with bi-level routing attention module (BiFormer) in the backbone. The BiFormer finds attention regions in the embryo image scene to obtain more edge and texture information of both pronuclei and blastomeres, which allows for a better feature realization of the neck region fusion interaction. Subsequently, to enhance model performance and mitigate redundant computation. The localization network uses partial convolution (PConv) in the backbone. The backbone network allows more efficient extraction of features by simultaneously reducing redundancy in computation with the effect of PConv. In addition, to mitigate the impact of low-quality samples in embryo images on localization, as well as to pay more attention to ordinary quality samples, we use wise intersection over union version 3 (WIoUv3) loss function with a dynamic non-monotonic focusing mechanism in the localization network, thus improving the overall performance of the algorithm. The experimental results show our model mAP@0.5 is 92.4% in localizing pronuclei and blastomeres. In practical terms, the ability to accurately localize pronuclei and blastomeres allows for better assessment of embryo quality and selects the best embryos.

Reparameterized underwater object detection network improved by cone-rod cell module and WIOU loss

To overcome the challenges in underwater object detection across diverse marine environments—marked by intricate lighting, small object presence, and camouflage—we propose an innovative solution inspired by the human retina's structure. This approach integrates a cone-rod cell module to counteract complex lighting effects and introduces a reparameterized multiscale module for precise small object feature extraction. Moreover, we employ the Wise Intersection Over Union (WIOU) technique to enhance camouflage detection. Our methodology simulates the human eye's cone and rod cells' brightness and color perception using varying sizes of deep and ordinary convolutional kernels. We further augment the network's learning capability and maintain model lightness through structural reparameterization, incorporating multi-branching and multiscale modules. By substituting the Complete Intersection Over Union (CIOU) with WIOU, we increase penalties for low-quality samples, mitigating the effect of camouflaged information on detection. Our model achieved a MAP_0.75 of 72.5% on the Real-World Underwater Object Detection (RUOD) dataset, surpassing the leading YOLOv8s model by 5.8%. Additionally, the model's FLOPs and parameters amount to only 10.62 M and 4.62B, respectively, which are lower than most benchmark models. The experimental outcomes affirm our design's efficacy in addressing underwater object detection's various disturbances, offering valuable technical insights for related oceanic image processing challenges.