Cureus. 2024 Mar 7;16(3):e55745. doi: 10.7759/cureus.55745. eCollection 2024 Mar.

ABSTRACT

This is a systematic review of 25 publications on the topics of the prevalence and cost of diabetic retinopathy (DR) in Trinidad and Tobago, the cost of traditional methods of screening for DR, and the use and cost of artificial intelligence (AI) in screening for DR. Analysis of these publications was used to identify and make estimates for how resources allocated to ophthalmology in public health systems in Trinidad and Tobago can be more efficiently utilized by employing AI in diagnosing treatable DR. DR screening was found to be an effective method of detecting the disease. Screening was found to be a universally cost-effective method of disease prevention and for altering the natural history of the disease in the spectrum of low-middle to high-income economies, such as Rwanda, Thailand, China, South Korea, and Singapore. AI and deep learning systems were found to be clinically superior to, or as effective as, human graders in areas where they were deployed, indicating that the systems are clinically safe. They have been shown to improve access to diabetic retinal screening, improve compliance with screening appointments, and prove to be cost-effective, especially in rural areas. Trinidad and Tobago, which is estimated to be disproportionately more affected by the burden of DR when projected out to the mid-21st century, stands to save as much as US$60 million annually from the implementation of an AI-based system to screen for DR versus conventional manual grading.

PMID:38586698 | PMC:PMC10999161 | DOI:10.7759/cureus.55745

Genet Res (Camb). 2024 Mar 30;2024:5549592. doi: 10.1155/2024/5549592. eCollection 2024.

ABSTRACT

22q11.2 deletion syndrome (22q11.2DS) is a microdeletion syndrome with a broad and heterogeneous phenotype, even though most of the deletions present similar sizes, involving ∼3 Mb of DNA. In a relatively large population of a Brazilian 22q11.2DS cohort (60 patients), we investigated genetic variants that could act as genetic modifiers and contribute to the phenotypic heterogeneity, using a targeted NGS (Next Generation Sequencing) with a specific Ion AmpliSeq panel to sequence nine candidate genes (CRKL, MAPK1, HIRA, TANGO2, PI4KA, HDAC1, ZDHHC8, ZFPM2, and JAM3), mapped in and outside the 22q11.2 hemizygous deleted region. In silico prediction was performed, and the whole-genome sequencing annotation analysis package (WGSA) was used to predict the possible pathogenic effect of single nucleotide variants (SNVs). For the in silico prediction of the indels, we used the genomic variants filtered by a deep learning model in NGS (GARFIELD-NGS). We identified six variants, 4 SNVs and 2 indels, in MAPK1, JAM3, and ZFPM2 genes with possibly synergistic deleterious effects in the context of the 22q11.2 deletion. Our results provide the opportunity for the discovery of the co-occurrence of genetic variants with 22q11.2 deletions, which may influence the patients´ phenotype.

PMID:38586596 | PMC:PMC10998724 | DOI:10.1155/2024/5549592

Front Genet. 2024 Mar 22;15:1377238. doi: 10.3389/fgene.2024.1377238. eCollection 2024.

ABSTRACT

The functional performance of immune cells relies on a complex transcriptional regulatory network. The three-dimensional structure of chromatin can affect chromatin status and gene expression patterns, and plays an important regulatory role in gene transcription. Currently available techniques for studying chromatin spatial structure include chromatin conformation capture techniques and their derivatives, chromatin accessibility sequencing techniques, and others. Additionally, the recently emerged deep learning technology can be utilized as a tool to enhance the analysis of data. In this review, we elucidate the definition and significance of the three-dimensional chromatin structure, summarize the technologies available for studying it, and describe the research progress on the chromatin spatial structure of dendritic cells, macrophages, T cells, B cells, and neutrophils.

PMID:38586584 | PMC:PMC10995239 | DOI:10.3389/fgene.2024.1377238

Heliyon. 2024 Mar 25;10(7):e28567. doi: 10.1016/j.heliyon.2024.e28567. eCollection 2024 Apr 15.

ABSTRACT

With the aging of the population, the quality of life and happiness of the elderly are increasingly becoming concerns of society. Smart senior care (SSC) products are an important tool to improve the quality of life of the elderly, and their application has been widely discussed. However, due to differences in cognitive characteristics and habits of the elderly, they may face difficulties when using smart products. Therefore, how to design a suitable SSC product APP interface for the elderly has become an urgent problem to be solved. The rapid development of deep learning (DL) provides a new way to deal with this problem. This paper aims to optimize the APP interface design of SSC products suitable for the elderly by using DL technology. Specifically, this paper designs a model based on a deep-Q-network (DQN) algorithm, which can better adapt to the usage habits and needs of the elderly through training and optimization. At the same time, the proposed model is evaluated comprehensively to verify its performance and superiority. To achieve the above goals, this paper first summarizes the existing SSC technology and the design suitable for the elderly. On this basis, a model based on the DQN algorithm is designed, and four datasets are used for training and testing. In the design of the model, this paper pays special attention to how to make the interface more friendly and easy to operate to meet the specific demands and preferences of the elderly. After a comprehensive evaluation, it is found that the proposed DQN algorithm model has achieved remarkable improvement in performance. This model has an average return of about 8.8 and 3.9 compared to other models, which have an average return of about 0.9 and 0.2, respectively. This shows that the model designed in this paper performs well in terms of accuracy and average return and performs better than other algorithms. The results of this paper reveal that optimizing the age-appropriate design of the SSC product APP interface through DL technology can notably enhance the user experience and satisfaction of the elderly. This not only helps the elderly to make better use of smart products and improve their quality of life but also provides useful inspiration and guidance for research and practice in related fields.

PMID:38586414 | PMC:PMC10998104 | DOI:10.1016/j.heliyon.2024.e28567

Heliyon. 2024 Mar 22;10(7):e28318. doi: 10.1016/j.heliyon.2024.e28318. eCollection 2024 Apr 15.

ABSTRACT

Urban expansion simulation is of significant importance to land management and policymaking. Advances in deep learning facilitate capturing and anticipating urban land dynamics with state-of-the-art accuracy properties. In this context, a novel deep learning-based ensemble framework was proposed for urban expansion simulation at an intra-urban granular level. The ensemble framework comprises i) multiple deep learning models as encoders, using transformers for encoding multi-temporal spatial features and convolutional layers for processing single-temporal spatial features, ii) a tailored channel-wise attention module to address the challenge of limited interpretability in deep learning methods. The channel attention module enables the examination of the rationality of feature importance, thereby establishing confidence in the simulated results. The proposed method accurately anticipated urban expansion in Shenzhen, China, and it outperformed all the baseline methods in terms of both spatial accuracy and temporal consistency.

PMID:38586370 | PMC:PMC10998072 | DOI:10.1016/j.heliyon.2024.e28318

Heliyon. 2024 Mar 27;10(7):e28270. doi: 10.1016/j.heliyon.2024.e28270. eCollection 2024 Apr 15.

ABSTRACT

In the face of environmental degradation and diminished energy resources, there is an urgent need for clean, affordable, and sustainable energy solutions, which highlights the importance of wind energy. In the global transition to renewable energy sources, wind power has emerged as a key player that is in line with the Paris Agreement, the Net Zero Target by 2050, and the UN 2030 Goals, especially SDG-7. It is critical to consider the variable and intermittent nature of wind to efficiently harness wind energy and evaluate its potential. Nonetheless, since wind energy is inherently variable and intermittent, a comprehensive assessment of a prospective site's wind power generation potential is required. This analysis is crucial for stakeholders and policymakers to make well-informed decisions because it helps them assess financial risks and choose the best locations for wind power plant installations. In this study, we introduce a framework based on Copula-Deep Learning within the context of decision trees. The main objective is to enhance the assessment of the wind power potential of a site by exploiting the intricate and non-linear dependencies among meteorological variables through the fusion of copulas and deep learning techniques. An empirical study was carried out using wind power plant data from Turkey. This dataset includes hourly power output measurements as well as comprehensive meteorological data for 2021. The results show that acknowledging and addressing the non-independence of variables through innovative frameworks like the Copula-LSTM based decision tree approach can significantly improve the accuracy and reliability of wind power plant potential assessment and analysis in other real-world data scenarios. The implications of this research extend beyond wind energy to inform decision-making processes critical for a sustainable energy future.

PMID:38586341 | PMC:PMC10998065 | DOI:10.1016/j.heliyon.2024.e28270

Plant Phenomics. 2024 Apr 5;6:0163. doi: 10.34133/plantphenomics.0163. eCollection 2024.

ABSTRACT

Asian soybean rust (ASR) is one of the major diseases that causes serious yield loss worldwide, even up to 80%. Early and accurate detection of ASR is critical to reduce economic losses. Hyperspectral imaging, combined with deep learning, has already been proved as a powerful tool to detect crop diseases. However, current deep learning models are limited to extract both spatial and spectral features in hyperspectral images due to the use of fixed geometric structure of the convolutional kernels, leading to the fact that the detection accuracy of current models remains further improvement. In this study, we proposed a deformable convolution and dilated convolution neural network (DC2Net) for the ASR detection. The deformable convolution module was used to extract the spatial features, while the dilated convolution module was applied to extract features from the spectral dimension. We also adopted the Shapley value and the channel attention methods to evaluate the importance of each wavelength during decision-making, thereby identifying the most contributing ones. The proposed DC2Net can realize early asymptomatic detection of ASR even when visual symptoms have not appeared. The results of the experiment showed that the detection performance of DC2Net dominated state-of-the-art methods, reaching an overall accuracy at 96.73%. Meanwhile, the experimental result suggested that the Shapley Additive exPlanations method was able to extract feature wavelengths correctly, thereby helping DC2Net achieve reasonable performance with less input data. The research result of this study could provide early warning of ASR outbreak in advance, even at the asymptomatic period.

PMID:38586218 | PMC:PMC10997487 | DOI:10.34133/plantphenomics.0163

Data Brief. 2024 Mar 21;54:110355. doi: 10.1016/j.dib.2024.110355. eCollection 2024 Jun.

ABSTRACT

This paper introduces an online dataset focused on detecting hairiness in yarn, including loop and protruding fibers. The dataset is designed for use in assessing artificial intelligence algorithms. The dataset consists of 684 original images. Through augmentation, this number increases to 1644, with 11,037 annotations derived from videos featuring 56.4tex purple cotton yarn. The videos were captured during the winding and unwinding processes of the purple yarn coil. An image acquisition system capable of capturing high-resolution images while the yarn is in motion was used, reaching speeds of up to 4.2 m/s and producing images with a resolution of 1.6M pixels. This dataset containing 100m of purple cotton yarn images was recorded and is available for download in various formats, including, among others, YOLOv8, YOLOv5, YOLOv7, MT-YOLOv6, COCO JSON, YOLO Darknet, Pascal VOC XML, TFRecord, CreateML JSON. Within an interface developed for a designed mechatronic prototype, users can choose to gather images or videos of yarn. Various characteristics of the yarn, such us: diameter, linear mass, volume, twist orientation, twist step, number of cables, hairiness index, number of loose fibers, thin places, thick places, neps (mass parameters) and U, CV and sH (statistical parameters) can be obtained. Recently, this online yarn spinning dataset was employed to validate artificial neural network models for real-time detection of hairiness in yarns, including loop fibers and protruding fibers. The dataset presented, with its clear annotations and wide array of augmentation techniques, serves as a foundational resource for prospective studies in textile engineering, enabling progress in the analysis and comprehension of yarn analysis.

PMID:38586143 | PMC:PMC10998032 | DOI:10.1016/j.dib.2024.110355

Data Brief. 2024 Mar 16;54:110334. doi: 10.1016/j.dib.2024.110334. eCollection 2024 Jun.

ABSTRACT

The Burkholderia glumae bacterium causes bacterial grain rot in rice, posing significant threats to the crop's yield, particularly thriving during the rice flowering and grain filling stages. This disease is especially evident in rice grains before harvest, presenting challenges in the detection and classification of rice panicles. Firstly, diseased grains may mix with healthy ones, complicating their separation. Secondly, the size of grains on a panicle varies from small to large, which can be problematic when detected using object detection methods. Thirdly, disease classification can be conducted by evaluating the extent of infection on rice panicles to assess its impact on yield. Finally, the challenges in detection, classification, and preprocessing for disease identification and management necessitate the adoption of diverse approaches in machine learning and deep learning to develop optimal methods and support smart agriculture.

PMID:38586139 | PMC:PMC10998030 | DOI:10.1016/j.dib.2024.110334

Pharmacoeconomics. 2024 Apr 7. doi: 10.1007/s40273-024-01371-1. Online ahead of print.

ABSTRACT

OBJECTIVES: Accurate risk stratification of patients with stage II and III colorectal cancer (CRC) prior to treatment selection enables limited health resources to be efficiently allocated to patients who are likely to benefit from adjuvant chemotherapy. We aimed to investigate the cost-effectiveness of a recently developed deep learning-based prognostic method, Histotyping, from the perspective of the Norwegian healthcare system.

METHODS: Two partitioned survival models were developed to assess the cost-effectiveness of Histotyping for two treatment cohorts: patients with CRC stage II and III. For each of the two cohorts, Histotyping was used for risk stratification to assign adjuvant chemotherapy and was compared with the standard of care (SOC) (adjuvant chemotherapy to all patients). Health outcomes measured in the model were quality-adjusted life years (QALYs) and life years (LYs) gained. Deterministic and probabilistic sensitivity analyses were performed to determine the impact of uncertainty. Scenario analyses were performed to assess the impact of the parameters with the greatest uncertainty.

RESULTS: Risk-stratifying patients with CRC stage II and III using Histotyping was dominant (less costly and more effective) compared to SOC. In patients with CRC stage II, the net monetary benefit of Histotyping was 270,934 Norwegian kroners (NOK) (year of valuation is 2021), and the net health benefit of Histotyping was 0.99. In stage III, the net monetary benefit of Histotyping was 195,419 NOK, and the net health benefit of Histotyping was 0.71.

CONCLUSIONS: Risk-stratifying patients with CRC using Histotyping prior to the administration of adjuvant chemotherapy is likely to be a cost-effective strategy in Norway.

PMID:38584239 | DOI:10.1007/s40273-024-01371-1

Front Optoelectron. 2024 Apr 8;17(1):9. doi: 10.1007/s12200-024-00112-8.

ABSTRACT

This paper presents an efficient scheme for single-pixel imaging (SPI) utilizing a phase-controlled fiber laser array and an untrained deep neural network. The fiber lasers are arranged in a compact hexagonal structure and coherently combined to generate illuminating light fields. Through the utilization of high-speed electro-optic modulators in each individual fiber laser module, the randomly modulated fiber laser array enables rapid speckle projection onto the object of interest. Furthermore, the untrained deep neural network is incorporated into the image reconstructing process to enhance the quality of the reconstructed images. Through simulations and experiments, we validate the feasibility of the proposed method and successfully achieve high-quality SPI utilizing the coherent fiber laser array at a sampling ratio of 1.6%. Given its potential for high emitting power and rapid modulation, the SPI scheme based on the fiber laser array holds promise for broad applications in remote sensing and other applicable fields.

PMID:38584213 | DOI:10.1007/s12200-024-00112-8

Spine J. 2024 Apr 5:S1529-9430(24)00159-1. doi: 10.1016/j.spinee.2024.03.019. Online ahead of print.

ABSTRACT

BACKGROUND CONTEXT: Adolescent idiopathic scoliosis (AIS) necessitates accurate spinal curvature assessment for effective clinical management. Traditional two-dimensional (2D) Cobb angle measurements have been the standard, but the emergence of three-dimensional (3D) automatic measurement techniques, such as those using weight-bearing 3D imaging (WR3D), presents an opportunity to enhance the accuracy and comprehensiveness of AIS evaluation.

PURPOSE: This study aimed to compare traditional 2D Cobb angle measurements with 3D automatic measurements utilizing the WR3D imaging technique in patients with AIS.

STUDY DESIGN/SETTING: A cohort of 53 AIS patients was recruited, encompassing 88 spinal curves, for comparative analysis.

PATIENT SAMPLE: The patient sample consisted of 53 individuals diagnosed with AIS.

OUTCOME MEASURES: Cobb angles were calculated using the conventional 2D method and three different 3D methods: the Analytical Method (AM), the Plane Intersecting Method (PIM), and the Plane Projection Method (PPM).

METHODS: The 2D cobb angle was manually measured by 3 experienced clinicians with 2D frontal whole-spine radiographs. For 3D cobb angle measurements, the spine and femoral heads were segmented from the WR3D images using a U-net deep-learning model, and the automatic calculations of the angles were performed with the 3D slicer software.

RESULTS: AM and PIM estimates were found to be significantly larger than 2D measurements. Conversely, PPM results showed no statistical difference compared to the 2D method. These findings were consistent in a subgroup analysis based on 2D Cobb angles.

CONCLUSION: Each 3D measurement method provides a unique assessment of spinal curvature, with PPM offering values closely resembling 2D measurements, while AM and PIM yield larger estimations. The utilization of WR3D technology alongside deep learning segmentation ensures accuracy and efficiency in comparative analyses. However, additional studies, particularly involving patients with severe curves, are required to validate and expand on these results. This study emphasizes the importance of selecting an appropriate measurement method considering the imaging modality and clinical context when assessing AIS, and it also underlines the need for continuous refinement of these techniques for optimal use in clinical decision-making and patient management.

PMID:38583576 | DOI:10.1016/j.spinee.2024.03.019

Gastrointest Endosc. 2024 Apr 5:S0016-5107(24)00213-X. doi: 10.1016/j.gie.2024.03.031. Online ahead of print.

ABSTRACT

BACKGROUND AND AIMS: The duodenal papillae are the primary and essential pathway for ERCP, greatly determining its complexity and outcome. We aimed to investigate the association between papilla morphology and post-ERCP pancreatitis (PEP), and to construct a robust model for PEP prediction.

METHODS: We enrolled retrospectively patients underwent ERCP in 2 centers from January 2019 and June 2022. Radiomic features of papilla were extracted from endoscopic images with deep learning. Potential predictors and their importance were evaluated with three machine learning algorithms. A predictive model was developed using best subset selection by logistic regression, and its performance was evaluated in terms of discrimination, calibration, and clinical utility based on area under curve (AUC) of receiver operation characteristics (ROC), calibration and clinical decision curve, respectively.

RESULTS: A total of 2038 and 334 ERCP patients from 2 centers were enrolled in this study with PEP rates of 7.9% and 9.6%, respectively. The R-score was significantly associated with PEP and showed great diagnostic value (AUC, 0.755-0.821). Six hub predictors were selected to conduct a predictive model. The radiomics-based model demonstrated excellent discrimination (AUC, 0.825-0.857) and therapeutic benefits in the training, testing, and validation cohorts. The addition of the R-score significantly improved diagnostic accuracy of the predictive model (NRI, 0.151-0.583, p<0.05; IDI, 0.097-0.235, p<0.001).

CONCLUSIONS: Radiomic signature of papilla is a crucial independent predictor of PEP. The papilla-radiomics-based model performs well for the clinical prediction of PEP.

PMID:38583542 | DOI:10.1016/j.gie.2024.03.031

Clin Neurophysiol. 2024 Mar 15;162:68-76. doi: 10.1016/j.clinph.2024.03.007. Online ahead of print.

ABSTRACT

OBJECTIVE: To evaluate the utility of a fully automated deep learning -based quantitative measure of EEG background, Brain State of the Newborn (BSN), for early prediction of clinical outcome at four years of age.

METHODS: The EEG monitoring data from eighty consecutive newborns was analyzed using the automatically computed BSN trend. BSN levels during the first days of life (a of total 5427 hours) were compared to four clinical outcome categories: favorable, cerebral palsy (CP), CP with epilepsy, and death. The time dependent changes in BSN-based prediction for different outcomes were assessed by positive/negative predictive value (PPV/NPV) and by estimating the area under the receiver operating characteristic curve (AUC).

RESULTS: The BSN values were closely aligned with four visually determined EEG categories (p < 0·001), as well as with respect to clinical milestones of EEG recovery in perinatal Hypoxic Ischemic Encephalopathy (HIE; p < 0·003). Favorable outcome was related to a rapid recovery of the BSN trend, while worse outcomes related to a slow BSN recovery. Outcome predictions with BSN were accurate from 6 to 48 hours of age: For the favorable outcome, the AUC ranged from 95 to 99% (peak at 12 hours), and for the poor outcome the AUC ranged from 96 to 99% (peak at 12 hours). The optimal BSN levels for each PPV/NPV estimate changed substantially during the first 48 hours, ranging from 20 to 80.

CONCLUSIONS: We show that the BSN provides an automated, objective, and continuous measure of brain activity in newborns.

SIGNIFICANCE: The BSN trend discloses the dynamic nature that exists in both cerebral recovery and outcome prediction, supports individualized patient care, rapid stratification and early prognosis.

PMID:38583406 | DOI:10.1016/j.clinph.2024.03.007

Artif Intell Med. 2024 Apr 1;151:102858. doi: 10.1016/j.artmed.2024.102858. Online ahead of print.

ABSTRACT

The unpredictable pandemic came to light at the end of December 2019, known as the novel coronavirus, also termed COVID-19, identified by the World Health Organization (WHO). The virus first originated in Wuhan (China) and rapidly affected most of the world's population. This outbreak's impact is experienced worldwide because it causes high mortality risk, many cases, and economic falls. Around the globe, the total number of cases and deaths reported till November 12, 2022, were >600 million and 6.6 million, respectively. During the period of COVID-19, several diverse diagnostic techniques have been proposed. This work presents a systematic review of COVID-19 diagnostic techniques in response to such acts. Initially, these techniques are classified into different categories based on their working principle and detection modalities, i.e. chest X-ray imaging, cough sound or respiratory patterns, RT-PCR, antigen testing, and antibody testing. After that, a comparative analysis is performed to evaluate these techniques' efficacy which may help to determine an optimum solution for a particular scenario. The findings of the proposed work show that Artificial Intelligence plays a vital role in developing COVID-19 diagnostic techniques which support the healthcare system. The related work can be a footprint for all the researchers, available under a single umbrella. Additionally, all the techniques are long-lasting and can be used for future pandemics.

PMID:38583369 | DOI:10.1016/j.artmed.2024.102858

Comput Methods Programs Biomed. 2024 Apr 2;249:108159. doi: 10.1016/j.cmpb.2024.108159. Online ahead of print.

ABSTRACT

BACKGROUND AND OBJECTIVE: Colorectal cancer (CRC) is one of the most commonly diagnosed cancers worldwide. The accurate survival prediction for CRC patients plays a significant role in the formulation of treatment strategies. Recently, machine learning and deep learning approaches have been increasingly applied in cancer survival prediction. However, most existing methods inadequately represent and leverage the dependencies among features and fail to sufficiently mine and utilize the comorbidity patterns of CRC. To address these issues, we propose a self-attention-based graph learning (SAGL) framework to improve the postoperative cancer-specific survival prediction for CRC patients.

METHODS: We present a novel method for constructing dependency graph (DG) to reflect two types of dependencies including comorbidity-comorbidity dependencies and the dependencies between features related to patient characteristics and cancer treatments. This graph is subsequently refined by a disease comorbidity network, which offers a holistic view of comorbidity patterns of CRC. A DG-guided self-attention mechanism is proposed to unearth novel dependencies beyond what DG offers, thus augmenting CRC survival prediction. Finally, each patient will be represented, and these representations will be used for survival prediction.

RESULTS: The experimental results show that SAGL outperforms state-of-the-art methods on a real-world dataset, with the receiver operating characteristic curve for 3- and 5-year survival prediction achieving 0.849±0.002 and 0.895±0.005, respectively. In addition, the comparison results with different graph neural network-based variants demonstrate the advantages of our DG-guided self-attention graph learning framework.

CONCLUSIONS: Our study reveals that the potential of the DG-guided self-attention in optimizing feature graph learning which can improve the performance of CRC survival prediction.

PMID:38583291 | DOI:10.1016/j.cmpb.2024.108159

Comput Methods Programs Biomed. 2024 Apr 3;249:108160. doi: 10.1016/j.cmpb.2024.108160. Online ahead of print.

ABSTRACT

BACKGROUND AND OBJECTIVE: Early detection and grading of Diabetic Retinopathy (DR) is essential to determine an adequate treatment and prevent severe vision loss. However, the manual analysis of fundus images is time consuming and DR screening programs are challenged by the availability of human graders. Current automatic approaches for DR grading attempt the joint detection of all signs at the same time. However, the classification can be optimized if red lesions and bright lesions are independently processed since the task gets divided and simplified. Furthermore, clinicians would greatly benefit from explainable artificial intelligence (XAI) to support the automatic model predictions, especially when the type of lesion is specified. As a novelty, we propose an end-to-end deep learning framework for automatic DR grading (5 severity degrees) based on separating the attention of the dark structures from the bright structures of the retina. As the main contribution, this approach allowed us to generate independent interpretable attention maps for red lesions, such as microaneurysms and hemorrhages, and bright lesions, such as hard exudates, while using image-level labels only.

METHODS: Our approach is based on a novel attention mechanism which focuses separately on the dark and the bright structures of the retina by performing a previous image decomposition. This mechanism can be seen as a XAI approach which generates independent attention maps for red lesions and bright lesions. The framework includes an image quality assessment stage and deep learning-related techniques, such as data augmentation, transfer learning and fine-tuning. We used the architecture Xception as a feature extractor and the focal loss function to deal with data imbalance.

RESULTS: The Kaggle DR detection dataset was used for method development and validation. The proposed approach achieved 83.7 % accuracy and a Quadratic Weighted Kappa of 0.78 to classify DR among 5 severity degrees, which outperforms several state-of-the-art approaches. Nevertheless, the main result of this work is the generated attention maps, which reveal the pathological regions on the image distinguishing the red lesions and the bright lesions. These maps provide explainability to the model predictions.

CONCLUSIONS: Our results suggest that our framework is effective to automatically grade DR. The separate attention approach has proven useful for optimizing the classification. On top of that, the obtained attention maps facilitate visual interpretation for clinicians. Therefore, the proposed method could be a diagnostic aid for the early detection and grading of DR.

PMID:38583290 | DOI:10.1016/j.cmpb.2024.108160

Mol Ther. 2024 Apr 5:S1525-0016(24)00219-3. doi: 10.1016/j.ymthe.2024.04.003. Online ahead of print.

ABSTRACT

Deep learning-based methods for protein structure prediction have achieved unprecedented accuracy. Yet, their utility in the engineering of protein-based binders remains constrained due to a gap between the ability to predict the structures of candidate proteins and the ability to assess which of those proteins are more probable to bind to a target. To bridge this gap, we introduce Automated Pairwise Peptide-Receptor AnalysIs for Screening Engineered proteins (APPRAISE), a method for predicting the target binding propensity of engineered proteins. After generating models of engineered proteins competing for binding to a target using an established structure-prediction tool such as AlphaFold-Multimer or ESMFold, APPRAISE performs a rapid (under 1 CPU second per model) scoring analysis that takes into account biophysical and geometrical constraints. As proof-of-concept cases, we demonstrate that APPRAISE can accurately classify receptor-dependent vs. receptor-independent adeno-associated viral vectors and diverse classes of engineered proteins such as miniproteins targeting the SARS-CoV-2 spike, nanobodies targeting a G-protein-coupled receptor, and peptides that specifically bind to transferrin receptor or PD-L1. APPRAISE is accessible through a web-based notebook interface using Google Colaboratory (https://tiny.cc/APPRAISE). With its accuracy, interpretability, and generalizability, APPRAISE promises to expand the utility of protein structure prediction and accelerate protein engineering for biomedical applications.

PMID:38582966 | DOI:10.1016/j.ymthe.2024.04.003

Sci Rep. 2024 Apr 6;14(1):8106. doi: 10.1038/s41598-024-58638-y.

ABSTRACT

Wheat head detection and counting using deep learning techniques has gained considerable attention in precision agriculture applications such as wheat growth monitoring, yield estimation, and resource allocation. However, the accurate detection of small and dense wheat heads remains challenging due to the inherent variations in their size, orientation, appearance, aspect ratios, density, and the complexity of imaging conditions. To address these challenges, we propose a novel approach called the Oriented Feature Pyramid Network (OFPN) that focuses on detecting rotated wheat heads by utilizing oriented bounding boxes. In order to facilitate the development and evaluation of our proposed method, we introduce a novel dataset named the Rotated Global Wheat Head Dataset (RGWHD). This dataset is constructed by manually annotating images from the Global Wheat Head Detection (GWHD) dataset with oriented bounding boxes. Furthermore, we incorporate a Path-aggregation and Balanced Feature Pyramid Network into our architecture to effectively extract both semantic and positional information from the input images. This is achieved by leveraging feature fusion techniques at multiple scales, enhancing the detection capabilities for small wheat heads. To improve the localization and detection accuracy of dense and overlapping wheat heads, we employ the Soft-NMS algorithm to filter the proposed bounding boxes. Experimental results indicate the superior performance of the OFPN model, achieving a remarkable mean average precision of 85.77% in oriented wheat head detection, surpassing six other state-of-the-art models. Moreover, we observe a substantial improvement in the accuracy of wheat head counting, with an accuracy of 93.97%. This represents an increase of 3.12% compared to the Faster R-CNN method. Both qualitative and quantitative results demonstrate the effectiveness of the proposed OFPN model in accurately localizing and counting wheat heads within various challenging scenarios.

PMID:38582913 | DOI:10.1038/s41598-024-58638-y

BMC Oral Health. 2024 Apr 6;24(1):426. doi: 10.1186/s12903-024-04160-6.

ABSTRACT

BACKGROUND: Dental development assessment is an important factor in dental age estimation and dental maturity evaluation. This study aimed to develop and evaluate the performance of an automated dental development staging system based on Demirjian's method using deep learning.

METHODS: The study included 5133 anonymous panoramic radiographs obtained from the Department of Pediatric Dentistry database at Seoul National University Dental Hospital between 2020 and 2021. The proposed methodology involves a three-step procedure for dental staging: detection, segmentation, and classification. The panoramic data were randomly divided into training and validating sets (8:2), and YOLOv5, U-Net, and EfficientNet were trained and employed for each stage. The models' performance, along with the Grad-CAM analysis of EfficientNet, was evaluated.

RESULTS: The mean average precision (mAP) was 0.995 for detection, and the segmentation achieved an accuracy of 0.978. The classification performance showed F1 scores of 69.23, 80.67, 84.97, and 90.81 for the Incisor, Canine, Premolar, and Molar models, respectively. In the Grad-CAM analysis, the classification model focused on the apical portion of the developing tooth, a crucial feature for staging according to Demirjian's method.

CONCLUSIONS: These results indicate that the proposed deep learning approach for automated dental staging can serve as a supportive tool for dentists, facilitating rapid and objective dental age estimation and dental maturity evaluation.

PMID:38582843 | DOI:10.1186/s12903-024-04160-6