Front Plant Sci. 2024 Oct 31;15:1398277. doi: 10.3389/fpls.2024.1398277. eCollection 2024.

ABSTRACT

INTRODUCTION: In response to the current mainstream deep learning detection methods with a large number of learned parameters and the complexity of apple leaf disease scenarios, the paper proposes a lightweight method and names it LCGSC-YOLO. This method is based on the LCNet(A Lightweight CPU Convolutional Neural Network) and GSConv(Group Shuffle Convolution) module modified YOLO(You Only Look Once) framework.

METHODS: Firstly, the lightweight LCNet is utilized to reconstruct the backbone network, with the purpose of reducing the number of parameters and computations of the model. Secondly, the GSConv module and the VOVGSCSP (Slim-neck by GSConv) module are introduced in the neck network, which makes it possible to minimize the number of model parameters and computations while guaranteeing the fusion capability among the different feature layers. Finally, coordinate attention is embedded in the tail of the backbone and after each VOVGSCSP module to improve the problem of detection accuracy degradation issue caused by model lightweighting.

RESULTS: The experimental results show the LCGSC-YOLO can achieve an excellent detection performance with mean average precision of 95.5% and detection speed of 53 frames per second (FPS) on the mixed datasets of Plant Pathology 2021 (FGVC8) and AppleLeaf9.

DISCUSSION: The number of parameters and Floating Point Operations (FLOPs) of the LCGSC-YOLO are much less thanother related comparative experimental algorithms.

PMID:39544536 | PMC:PMC11560749 | DOI:10.3389/fpls.2024.1398277

Front Plant Sci. 2024 Oct 31;15:1460060. doi: 10.3389/fpls.2024.1460060. eCollection 2024.

ABSTRACT

With the advancement of computer vision technology, vision-based target perception has emerged as a predominant approach for harvesting robots to identify and locate fruits. However, little attention has been paid to the fact that fruits may be obscured by stems or other objects. In order to improve the vision detection ability of fruit harvesting robot, a fruit target selection and location approach considering obstacle perception was proposed. To enrich the dataset for tomato harvesting, synthetic data were generated by rendering a 3D simulated model of the tomato greenhouse environment, and automatically producing corresponding pixel-level semantic segmentation labels. An attention-based spatial-relationship feature extraction module (SFM) with lower computational complexity was designed to enhance the ability of semantic segmentation network DeepLab v3+ in accurately segmenting linear-structured obstructions such as stems and wires. An adaptive K-means clustering method was developed to distinguish individual instances of fruits. Furthermore, a barrier-free fruit selection algorithm that integrates information of obstacles and fruit instances was proposed to identify the closest and largest non-occluded fruit as the optimal picking target. The improved semantic segmentation network exhibited enhanced performance, achieving an accuracy of 96.75%. Notably, the Intersection-over-Union (IoU) of wire and stem classes was improved by 5.0% and 2.3%, respectively. Our target selection method demonstrated accurate identification of obstacle types (96.15%) and effectively excluding fruits obstructed by strongly resistant objects (86.67%). Compared to the fruit detection method without visual obstacle avoidance (Yolo v5), our approach exhibited an 18.9% increase in selection precision and a 1.3% reduction in location error. The improved semantic segmentation algorithm significantly increased the segmentation accuracy of linear-structured obstacles, and the obstacle perception algorithm effectively avoided occluded fruits. The proposed method demonstrated an appreciable ability in precisely selecting and locating barrier-free fruits within non-structural environments, especially avoiding fruits obscured by stems or wires. This approach provides a more reliable and practical solution for fruit selection and localization for harvesting robots, while also being applicable to other fruits and vegetables such as sweet peppers and kiwis.

PMID:39544532 | PMC:PMC11560766 | DOI:10.3389/fpls.2024.1460060

Health Inf Sci Syst. 2024 Nov 12;12(1):53. doi: 10.1007/s13755-024-00311-9. eCollection 2024 Dec.

ABSTRACT

Depression is one of the most prevalent mental conditions which could impair people's productivity and lead to severe consequences. The diagnosis of this disease is complex as it often relies on a physician's subjective interview-based screening. The aim of our work is to propose deep learning models for automatic depression detection by using different data modalities, which could assist in the diagnosis of depression. Current works on automatic depression detection mostly are tested on a single dataset, which might lack robustness, flexibility and scalability. To alleviate this problem, we design a novel Graph Neural Network-enhanced Transformer model named DePressionDetect Net (DPD Net) that leverages textual, audio and visual features and can work under two different application settings: the clinical setting and the social media setting. The model consists of a unimodal encoder module for encoding single modality, a multimodal encoder module for integrating the multimodal information, and a detection module for producing the final prediction. We also propose a model named DePressionDetect-with-EEG Net (DPD-E Net) to incorporate Electroencephalography (EEG) signals and speech data for depression detection. Experiments across four benchmark datasets show that DPD Net and DPD-E Net can outperform the state-of-the-art models on three datasets (i.e., E-DAIC dataset, Twitter depression dataset and MODMA dataset), and achieve competitive performance on the fourth one (i.e., D-vlog dataset). Ablation studies demonstrate the advantages of the proposed modules and the effectiveness of combining diverse modalities for automatic depression detection.

PMID:39544256 | PMC:PMC11557813 | DOI:10.1007/s13755-024-00311-9

J Biol Methods. 2024 Aug 9;11(3):e99010017. doi: 10.14440/jbm.2024.0016. eCollection 2024.

ABSTRACT

Gene expression data are used to discover meaningful hidden information in gene datasets. Cancer and other disorders may be diagnosed based on differences in gene expression profiles, and this information can be gleaned by gene sequencing. Thanks to the tremendous power of artificial intelligence (AI), healthcare has become a significant user of deep learning (DL) for predicting cancer diseases and categorizing gene expression. Gene expression Microarrays have been proved effective in predicting cancer diseases and categorizing gene expression. Gene expression datasets contain only limited samples, but the features of cancer are diverse and complex. To overcome their dimensionality, gene expression datasets must be enhanced. By learning and analyzing features of input data, it is possible to extract features, as multidimensional arrays, from the data. Synthetic samples are needed to strengthen the range of information. DL strategies may be used when gene expression data are used to diagnose and classify cancer diseases.

PMID:39544183 | PMC:PMC11557296 | DOI:10.14440/jbm.2024.0016

Can Assoc Radiol J. 2024 Nov 15:8465371241296834. doi: 10.1177/08465371241296834. Online ahead of print.

ABSTRACT

Purpose: Pediatric low-grade gliomas (pLGG) are the most common brain tumour in children, and the molecular diagnosis of pLGG enables targeted treatment. We use MRI-based Convolutional Neural Networks (CNNs) for molecular subtype identification of pLGG and augment the models using tumour location probability maps. Materials and Methods: MRI FLAIR sequences of 214 patients (110 male, mean age of 8.54 years, 143 BRAF fused and 71 BRAF V600E mutated pLGG tumours) from January 2000 to December 2018 were included in this retrospective REB-approved study. Tumour segmentations (volumes of interest-VOIs) were provided by a pediatric neuroradiology fellow and verified by a pediatric neuroradiologist. Patients were randomly split into development and test sets with an 80/20 ratio. The 3D binary VOI masks for each class in the development set were combined to derive the probability density functions of tumour location. Three pipelines for molecular diagnosis of pLGG were developed: location-based, CNN-based, and hybrid. The experiment was repeated 100 times each with different model initializations and data splits, and the Areas Under the Receiver Operating Characteristic Curve (AUROC) was calculated, and Student's t-test was conducted. Results: The location-based classifier achieved an AUROC of 77.9, 95% confidence interval (CI) (76.8, 79.0). CNN-based classifiers achieved an AUROC of 86.1, 95% CI (85.0, 87.3), while the tumour-location-guided CNNs outperformed the other classifiers with an average AUROC of 88.64, 95% CI (87.6, 89.7), which was statistically significant (P-value .0018). Conclusion: Incorporating tumour location probability maps into CNN models led to significant improvements for molecular subtype identification of pLGG.

PMID:39544176 | DOI:10.1177/08465371241296834

Anal Methods. 2024 Nov 15. doi: 10.1039/d4ay01346h. Online ahead of print.

ABSTRACT

Agriculture has a substantial demand for classification, and each agricultural product exhibits a unique ion signal. This paper summarizes the classification techniques of ion-selective electrode arrays in agriculture. Initially, data sample collection methods based on ion-selective electrode arrays are summarized. The paper then discusses the current state of classification algorithms from the perspectives of machine learning, artificial neural networks, extreme learning machines, and deep learning, along with their existing research in ion-selective electrodes and related fields. Then, the potential applications in crop and livestock growth status classification, soil classification, agricultural product quality classification, and agricultural product type classification are discussed. Ultimately, the future challenges of ion-selective electrode research are discussed from the perspectives of the sensor itself and algorithms combined with sensor arrays, which also positively impact the promotion of their application in agriculture. This work will advance the application of classification techniques combined with ion-selective electrode arrays in agriculture.

PMID:39543972 | DOI:10.1039/d4ay01346h

J Clin Pediatr Dent. 2024 Nov;48(6):161-172. doi: 10.22514/jocpd.2024.136. Epub 2024 Nov 3.

ABSTRACT

Traditional tooth image analysis methods primarily focus on feature extraction from individual images, often overlooking critical tooth shape and position information. This paper presents a novel computer-aided diagnosis method, Collaborative learning with Mask Region-based Convolutional Neural Network (Co-Mask R-CNN), designed to enhance tooth image analysis by leveraging the integration of complementary information. First, image enhancement is employed to generate an edge-enhanced tooth edge image. Then, a collaborative learning strategy combined with Mask R-CNN is introduced, where the original and edge images are input simultaneously, and a two-stream encoder extracts feature maps from complementary images. By utilizing an attention mechanism, the output features from the two branches are dynamically fused, quantifying the relative importance of the two complementary images at different spatial positions. Finally, the fused feature map is utilized for tooth instance segmentation. Extensive experiments are conducted using a proprietary dataset to evaluate the effectiveness of Co-Mask R-CNN, and the results are compared against those of an alternative segmentation network. The results demonstrate that Co-Mask R-CNN outperforms the other networks in terms of both segmentation accuracy and robustness. Consequently, this method holds considerable promise for providing medical professionals with precise tooth segmentation results, establishing a reliable foundation for subsequent tooth disease diagnosis and treatment.

PMID:39543893 | DOI:10.22514/jocpd.2024.136

Nat Comput Sci. 2024 Nov 14. doi: 10.1038/s43588-024-00725-1. Online ahead of print.

NO ABSTRACT

PMID:39543392 | DOI:10.1038/s43588-024-00725-1

Sci Rep. 2024 Nov 14;14(1):28019. doi: 10.1038/s41598-024-79175-8.

ABSTRACT

Efficient prediction of blastocyst formation from early-stage human embryos is imperative for improving the success rates of assisted reproductive technology (ART). Clinics transfer embryos at the blastocyst stage on Day-5 but Day-3 embryo transfer offers the advantage of a shorter culture duration, which reduces exposure to laboratory conditions, potentially enhancing embryonic development within a more conducive uterine environment and improving the likelihood of successful pregnancies. In this paper, we present a novel ResNet-GRU deep-learning model to predict blastocyst formation at 72 HPI. The model considers the time-lapse images from the incubator from Day 0 to Day 3. The model predicts blastocyst formation with a validation accuracy of 93% from the cleavage stage. The sensitivity and specificity are 0.97 and 0.77 respectively. The deep learning model presented in this paper will assist the embryologist in identifying the best embryo to transfer at Day 3, leading to improved patient outcomes and pregnancy rates in ART.

PMID:39543360 | DOI:10.1038/s41598-024-79175-8

Nat Methods. 2024 Nov 14. doi: 10.1038/s41592-024-02511-3. Online ahead of print.

ABSTRACT

The rapid growth of single-cell transcriptomic technology has produced an increasing number of datasets for both embryonic development and in vitro pluripotent stem cell-derived models. This avalanche of data surrounding pluripotency and the process of lineage specification has meant it has become increasingly difficult to define specific cell types or states in vivo, and compare these with in vitro differentiation. Here we utilize a set of deep learning tools to integrate and classify multiple datasets. This allows the definition of both mouse and human embryo cell types, lineages and states, thereby maximizing the information one can garner from these precious experimental resources. Our approaches are built on recent initiatives for large-scale human organ atlases, but here we focus on material that is difficult to obtain and process, spanning early mouse and human development. Using publicly available data for these stages, we test different deep learning approaches and develop a model to classify cell types in an unbiased fashion at the same time as defining the set of genes used by the model to identify lineages, cell types and states. We used our models trained on in vivo development to classify pluripotent stem cell models for both mouse and human development, showcasing the importance of this resource as a dynamic reference for early embryogenesis.

PMID:39543284 | DOI:10.1038/s41592-024-02511-3

Sci Rep. 2024 Nov 14;14(1):28050. doi: 10.1038/s41598-024-79111-w.

ABSTRACT

Parkinsonian tremor (PT) and Essential tremor (ET) exist as upper limb tremors in clinical practice. Notably, their types of trembling share similar presentations and overlapping frequencies. To enhance objectivity and efficiency in the diagnosis of these two diseases, there is a pressing need for more objective tremor classification procedures. This study proposes a novel multimodal fusion network based on a cross-attention mechanism (MFCA-Net) to automate the classification of upper limb tremors between PTs and ETs. To this end, 140 patients with PTs and ETs were recruited, and acceleration and surface electromyography (sEMG) signals were collected from the forearm during tremor episodes. To comprehensively capture the global and local features of input signals, a multiscale convolution in MFCA-Net was designed. Furthermore, the cross-attention mechanism was applied to fuse the features of the two input signals. The results demonstrate that the final classification accuracy exceeded 97.18% when MFCA-Net was used. Compared with the single acceleration signal and single sEMG signal inputs, the recognition accuracies increased by 18.91% and 10.04%, respectively. Therefore, the proposed MFCA-Net in this study serves as an objective and potential tool for assisting clinicians in the diagnosis of PT and ET patients.

PMID:39543261 | DOI:10.1038/s41598-024-79111-w

Sci Rep. 2024 Nov 14;14(1):28059. doi: 10.1038/s41598-024-78979-y.

ABSTRACT

The performance gains achieved by deep learning models nowadays are mainly attributed to the usage of ever larger datasets. In this study, we present and contrast the performance gains that can be achieved via accessing larger high-quality datasets versus the gains that can be achieved from harnessing the latest deep learning architectural and training advances. Modelling neonatal EEG is particularly affected by the lack of publicly available large datasets. It is shown that greater performance gains can be achieved from harnessing the latest deep learning advances than using a larger training dataset when adopting AUC as a metric, whereas using AUC90 or AUC-PR as metrics greater performance gains are achieved from using a larger dataset than harnessing the latest deep learning advances. In all scenarios the best performance is obtained by combining both deep learning advances and larger datasets. A novel developed architecture is presented that outperforms the current state-of-the-art model for the task of neonatal seizure detection. A novel method to fine-tune the presented model towards site-specific settings based on pseudo labelling is also outlined. The code and the weights of the model are made publicly available for benchmarking future model performances for neonatal seizure detection.

PMID:39543245 | DOI:10.1038/s41598-024-78979-y

Sci Rep. 2024 Nov 14;14(1):28063. doi: 10.1038/s41598-024-78635-5.

ABSTRACT

The rapid advancement of artificial intelligence and robot technology has spurred the proposal and innovation of a coal gangue sorting robot system (CGSRS) paradigm. The time-varying raw coal flow (TVRCF) with multi-scene and full working conditions affects the gangue queue. Configuring the CGSRS scheme correctly is combative. The field environment puts forward higher requirements for the time complexity of the CGSRS multi-task allocation strategy. Therefore, this paper proposes a scheme evaluation method of the CGSRS with time-varying multi-scenario based on deep learning. Firstly, the gangue queue data set of multi-scene and full-condition TVRCF was obtained according to the belt speed, the maximum coal flow, and the uncorrelated nonlinear changes of coal flow and gangue content. The CGSRS scheme is established based on robot number and rule combination, and the multi-task allocation strategy is adjusted to generate the labels of the gangue queue. Then, the RGB sample set is established based on the labels of the gangue queue. The CGSRS scheme evaluation model is trained based on DenseNet. Finally, the CGSRS scheme evaluation method was designed to realize the prediction of a random gangue queue. In this paper, the CGSRS scheme evaluation model, the stability of the solution, and the comparison of methods are carried out. Experimental results show that the solution of the CGSRS scheme evaluation model is accurate and stable. The time complexity is significantly reduced and very stable. The CGSRS scheme evaluation method is applied to the CGSRS multi-task allocation problem, and the stability of the solution is not affected by the data. It is significantly better than the multi-task allocation strategy. The proposed method is the first attempt to apply deep learning to a multi-task allocation problem in CGSRS.

PMID:39543191 | DOI:10.1038/s41598-024-78635-5

Sci Rep. 2024 Nov 14;14(1):28024. doi: 10.1038/s41598-024-77554-9.

ABSTRACT

Distributed denial of service (DDoS) attack is one of the most hazardous assaults in cloud computing or networking. By depleting resources, this attack renders the services unavailable to end users and leads to significant financial and reputational damage. Hence, identifying such threats is crucial to minimize revenue loss, market share, and productivity loss and enhance the brand reputation. In this study, we implemented an effective intrusion detection system using deep learning approach. The suggested framework includes three phases: Data pre-processing, Data balancing, and Classification. First, we prepare the valid data, which is helpful for further processing. Then, we balance the given pre-processed data by Conditional generative adversarial network (CGAN), and as a result, we can minimize the bias towards the majority classes. Finally, we distinguish whether the traffic is attack or benign using a stacked sparse denoising autoencoder (SSDAE) with a firefly-black widow (FA-BW) hybrid optimization algorithm. All these experiments are validated through the CICDDoS2019 dataset and compared with well-received techniques. From these findings, we observed that the proposed strategy detects DDoS attacks significantly more accurately than other approaches. Based on our findings, this study highlights the crucial role played by advanced deep learning techniques and hybrid optimization algorithms in strengthening cybersecurity against DDoS attacks.

PMID:39543174 | DOI:10.1038/s41598-024-77554-9

Nat Commun. 2024 Nov 14;15(1):9886. doi: 10.1038/s41467-024-54182-5.

ABSTRACT

Cancer is a heterogeneous disease requiring costly genetic profiling for better understanding and management. Recent advances in deep learning have enabled cost-effective predictions of genetic alterations from whole slide images (WSIs). While transformers have driven significant progress in non-medical domains, their application to WSIs lags behind due to high model complexity and limited dataset sizes. Here, we introduce SEQUOIA, a linearized transformer model that predicts cancer transcriptomic profiles from WSIs. SEQUOIA is developed using 7584 tumor samples across 16 cancer types, with its generalization capacity validated on two independent cohorts comprising 1368 tumors. Accurately predicted genes are associated with key cancer processes, including inflammatory response, cell cycles and metabolism. Further, we demonstrate the value of SEQUOIA in stratifying the risk of breast cancer recurrence and in resolving spatial gene expression at loco-regional levels. SEQUOIA hence deciphers clinically relevant information from WSIs, opening avenues for personalized cancer management.

PMID:39543087 | DOI:10.1038/s41467-024-54182-5

Methods Mol Biol. 2025;2870:341-355. doi: 10.1007/978-1-0716-4213-9_17.

ABSTRACT

Rapid advancements in cryogenic electron microscopy (cryo-EM) have revolutionized the field of structural biology by enabling the determination of complex macromolecular structures at unprecedented resolutions. When cryo-EM density maps have a resolution around 3 Å, the atomic structure can be modeled manually. However, as the resolution decreases, analyzing these density maps becomes increasingly challenging. For modeling structures in lower resolution maps, deep learning can be used to identify structural features in the maps to assist in structure modeling.Here, we present a suite of deep learning-based tools developed by our lab that enable structural biologists to work with cryo-EM maps of a wide range of resolutions. For cryo-EM maps at near-atomic resolution (5 Å or better), DeepMainmast automatically models all-atom structures by tracing the main chain from local map features of amino acids and atoms detected by deep learning; DAQ score quantifies map-model fit and indicates potential misassignments in protein models. In intermediate resolution maps (5-10 Å), Emap2sec and Emap2sec+ can accurately detect protein secondary structures and nucleic acids. These tools and more are available at our web server: https://em.kiharalab.org/ .

PMID:39543043 | DOI:10.1007/978-1-0716-4213-9_17

Methods Mol Biol. 2025;2870:1-19. doi: 10.1007/978-1-0716-4213-9_1.

ABSTRACT

The secondary structures (SSs) and supersecondary structures (SSSs) underlie the three-dimensional structure of proteins. Prediction of the SSs and SSSs from protein sequences enjoys high levels of use and finds numerous applications in the development of a broad range of other bioinformatics tools. Numerous sequence-based predictors of SS and SSS were developed and published in recent years. We survey and analyze 45 SS predictors that were released since 2018, focusing on their inputs, predictive models, scope of their prediction, and availability. We also review 32 sequence-based SSS predictors, which primarily focus on predicting coiled coils and beta-hairpins and which include five methods that were published since 2018. Substantial majority of these predictive tools rely on machine learning models, including a variety of deep neural network architectures. They also frequently use evolutionary sequence profiles. We discuss details of several modern SS and SSS predictors that are currently available to the users and which were published in higher impact venues.

PMID:39543027 | DOI:10.1007/978-1-0716-4213-9_1

Spectrochim Acta A Mol Biomol Spectrosc. 2024 Nov 10;327:125417. doi: 10.1016/j.saa.2024.125417. Online ahead of print.

ABSTRACT

Design of a sensitive, cost-effective SERS substrate is critical for probing analyte in trace concentration in real field environment. Present work reports the fabrication of an oxygen (O2) plasma treated bimetallic nanofibers as a sensitive SERS platform. In contrast to the conventional nanofiber-based SERS platform, the proposed plasma-treated bimetallic nanofibers-based SERS platform offers high sensitivity and reproducibility characteristics. On top, the use of bimetallic nanoparticles provides a synergistic effect, contributing to both electromagnetic and chemical enhancement to SERS performance and the plasma treatment contributes to the controlled exposure of the embedded nanoparticles (NPs) to the analyte thereby enhancing the overall sensitivity of the proposed technique. With standard Raman active probe molecules - 1,2-bis(4-pyridyl) ethylene (BPE) and rhodamine-6G (R6G) the limit of detection (LOD) and the limit of quantification (LOQ) of the proposed sensing platform are estimated to be 3.8 nM and 11.6 nM respectively. The enhancement factor (EF) of the designed sensing platform is calculated to be ∼108 with a maximum signal variations of 5 %. The applicability of the designed SERS substrate has been realized through detection of two antibiotics - fluconazole (FLU) and lincomycin (LIN) widely used in poultry farms. Furthermore, a deep learning model - artificial neural network (ANN) has been implemented for effective classification of the analyte molecules from a mixed sample.

PMID:39541643 | DOI:10.1016/j.saa.2024.125417

Int J Med Inform. 2024 Nov 13;193:105698. doi: 10.1016/j.ijmedinf.2024.105698. Online ahead of print.

ABSTRACT

INTRODUCTION: Ultrasound equipment provides real-time visualization of internal organs, essential for early disease detection and diagnosis. However, poor-quality ultrasound images can compromise diagnostic accuracy and increase the risk of misdiagnosis. Quality assessments are often subjective, relying on the evaluator's experience and interpretation, which can vary.

METHODS: This study introduces a two-stage deep learning framework designed to objectively assess ultrasound image quality using phantom data across three key parameters: 'Dead zone', 'Axial/lateral resolution', and 'Gray scale and dynamic range'. Stage 1 automatically extracts regions of interest for each parameter, while Stage 2 employs detection or classification models to evaluate image quality within these regions. To generate an overall equipment quality score, a logistic regression model combines the weighted results from each parameter.

RESULTS: The classification model demonstrated high performance across datasets, achieving AUC scores of 98.6% for 'Dead zone', 87.7% for 'Axial/lateral resolution', and 96.0% for 'Gray scale and dynamic range'. Further analysis using guideline-compliant images of individual devices showed AUC scores of 98.2%, 92.8%, and 100%, respectively. These findings highlight deep learning's potential for quantitative and objective assessments of ultrasound image quality. Ultimately, this framework provides a streamlined approach to quality management, enabling consistent quality control and efficient scoring-based evaluation of ultrasound equipment.

PMID:39541619 | DOI:10.1016/j.ijmedinf.2024.105698

PLoS One. 2024 Nov 14;19(11):e0313640. doi: 10.1371/journal.pone.0313640. eCollection 2024.

NO ABSTRACT

PMID:39541288 | DOI:10.1371/journal.pone.0313640