3D Convolutional Neural Network Model Research Articles

Defects recognition of pine nuts using hyperspectral imaging and deep learning approaches

Pine nuts, as a highly nutritious and medicinally valuable food, are susceptible to various defects during their cultivation, harvesting, and transportation, which can reduce their quality. Therefore, rapid and accurate identification of pine nut defect types is of utmost importance to ensure the overall quality of the pine nuts. In this study, hyperspectral imaging (HSI) systems covering two spectral ranges (400–1000 nm and 900–1700 nm) were employed to capture hyperspectral images of healthy pine nuts and pine nuts with six types of defects. One-dimensional (1D) and three-dimensional (3D) Convolutional Neural Network (CNN) models with multi-head attention mechanisms were constructed using 1D spectra and 3D hyperspectral images, respectively. To validate the effectiveness of the proposed models, Support Vector Classifier (SVC) models were built using 1D spectra and used as a comparison. Overall, the proposed CNN models outperform traditional machine learning methods in two spectral ranges (400–1000 nm and 900–1700 nm). 1D CNN model in the near-infrared spectral range (900–1700 nm) achieved an accuracy of 90.23 % on the training set and 81.32 % on the validation set. Additionally, the Generalized Gradient-Weighted Class Activation Mapping (Grad-CAM++) visualization method was applied to conduct visual analysis on the 1D CNN and 3D CNN models, enabling the identification of important wavelength ranges and pixel regions in the models, thereby enhancing the interpretability of the decision-making process of the models. Overall, the results of this study demonstrated the feasibility of using a combination of hyperspectral imaging and convolutional neural networks for pine nut defects classification, and the visual analysis of the models provided new insights and understanding for pine nut defects identification.

Advanced wind turbine blade inspection with hyperspectral imaging and 3D convolutional neural networks for damage detection

In the context of global efforts to mitigate climate change by pursuing sustainable energy sources, wind energy has emerged as a critical contributor. However, the wind energy industry faces substantial challenges in maintaining and preserving the integrity of wind turbine blades. Timely and accurate detection and classification of blade faults, encompassing issues such as cracks, erosion, and ice buildup, are imperative to uphold wind turbines' ongoing efficiency and safety. This study introduces an inventive approach that amalgamates hyperspectral imaging and 3D Convolutional Neural Networks (CNNs) to augment the precision and efficiency of wind turbine blade fault detection and classification. Hyperspectral imaging is harnessed to capture comprehensive spectral information from blade surfaces, facilitating exact fault identification. The process is streamlined through Incremental Principal Component Analysis (IPCA), reducing data dimensions while maintaining integrity. The 3D CNN model demonstrates remarkable performance, achieving high accuracy in detecting all fault categories in full-band hyperspectral images. The model retains high accuracy even with dimensionality reduction to 20 spectral bands. The reduced processing time of the 20-band image enhances the practicality of real-world applications, thereby reducing downtime and maintenance expenditures. This research represents a significant advancement in wind turbine blade inspection, contributing to the sustainability and dependability of wind energy systems and furthering the cause of a cleaner and more sustainable energy future as part of the broader fight against climate change.

Computational limits to the legibility of the imaged human brain

Our knowledge of the organisation of the human brain at the population-level is yet to translate into power to predict functional differences at the individual-level, limiting clinical applications and casting doubt on the generalisability of inferred mechanisms. It remains unknown whether the difficulty arises from the absence of individuating biological patterns within the brain, or from limited power to access them with the models and compute at our disposal.Here we comprehensively investigate the resolvability of such patterns with data and compute at unprecedented scale. Across 23 810 unique participants from UK Biobank, we systematically evaluate the predictability of 25 individual biological characteristics, from all available combinations of structural and functional neuroimaging data. Over 4526 GPU*hours of computation, we train, optimize, and evaluate out-of-sample 700 individual predictive models, including fully-connected feed-forward neural networks of demographic, psychological, serological, chronic disease, and functional connectivity characteristics, and both uni- and multi-modal 3D convolutional neural network models of macro- and micro-structural brain imaging.We find a marked discrepancy between the high predictability of sex (balanced accuracy 99.7%), age (mean absolute error 2.048 years, R2 0.859), and weight (mean absolute error 2.609Kg, R2 0.625), for which we set new state-of-the-art performance, and the surprisingly low predictability of other characteristics. Neither structural nor functional imaging predicted an individual's psychology better than the coincidence of common chronic disease (p < 0.05). Serology predicted chronic disease (p < 0.05) and was best predicted by it (p < 0.001), followed by structural neuroimaging (p < 0.05).Our findings suggest either more informative imaging or more powerful models will be needed to decipher individual level characteristics from the human brain. We make our models and code openly available.

3D mineral prospectivity modeling using multi-scale 3D convolution neural network and spatial attention approaches

A significant proportion of recent mineral exploration has increasingly focused on the targeting of deep-seated orebodies. Mineral prospectivity modeling is one of the more important approaches that facilitates exploration targeting and the mitigation of risks associated with mineral exploration, particularly under cover. Recent advances in 3D mineral prospectivity modeling enable the effective extraction of predictive information from three-dimensional geological models, enabling more accurate exploration targeting of deep-seated orebodies. These advancements have synergized with deep learning approaches to improve the efficiency of mineral exploration based on nonlinear and multi-layer sensing attributes, effectively enabling the identification and extraction of key relationships between 3D predictive maps and mineralization. The main deep learning method used for 3D mineral prospectivity modeling is convolutional neural network (CNN) modeling. However, this research typically does not consider the multiscale features of geological structures, meaning further improvements can be made to this modeling approach. This paper introduces a multi-scale 3D convolutional neural network model (3D CNN) incorporating a spatial attention mechanism and an Inception module (MSAM-CNN) for 3D mineral prospectivity modeling. By integrating Inception modules and spatial attention mechanisms, the network's capability to identify multi-scale geological features and delineate key predictive areas is significantly enhanced compared to typical CNN approaches. This new approach provides further improvement in the accuracy and generalization capability of 3D mineral prospectivity modeling. To evaluate the effectiveness of this model, we undertook 3D mineral prospectivity modeling within the area of the Baixiangshan iron deposit, in the Ningwu Basin of the Middle-Lower Yangtze River Metallogenic Belt, China. The results show that the multi-scale 3D convolutional neural network model is remarkably robust and has good generalization capabilities. The approach can also can effectively delineate targets within the deep and peripheral areas of the deposit, providing targets for future exploration. The addition, performance indicators, ROC curve, and Capture-Efficiency curve generated during this modeling consistently demonstrate that the MSAM-CNN model outperforms Inception-enhanced CNN (M-CNN), CNN, Random Forest (RF), and Support Vector Machine (SVM) models. All of this indicates that MSAM-CNN approaches can effectively extract 3D spatial features within 3D predictive maps during 3D mineral prospectivity modeling better than other approaches that are commonly used, indicating that thius approach represents a promising tool for the accurate and precise identification of targets during future exploration for deep-seated mineralization.

A hybrid convolution transformer for hyperspectral image classification

ABSTRACT Hyperspectral images play a crucial role in remote sensing applications surveillance, environment and precision agriculture, containing abundant object information. However, they often face challenges such as limited labelled data and imbalanced classes. In recent years, convolutional neural networks (CNNs) have shown impressive performance in computer vision tasks, including hyperspectral image classification. The emergence of transformers has also attracted attention for hyperspectral image analysis due to their promising capabilities. Nevertheless, transformers typically demand a substantial amount of training data, making their application challenging in scenarios with limited labelled samples. To overcome this limitation, we propose a hybrid convolution transformer framework. Our method uses a vision transformer and a residual 3D convolutional neural network model. It also uses a sequence aggregation layer to avoid overfitting issues that come up when there isn’t enough training data. Our proposed residual channel attention module captures richer spatial-spectral complementary information and maintains spectral details during the feature extraction process. We conducted experiments on three benchmark datasets. The proposed model achieved state of the art performance 99.75 % , 99.46 % and 99.95 % in terms of overall accuracy (OA) using only 5 % , 5 % and 10 % labelled training samples respectively. Which is better than other state of the art methods.

Three dimensional convolutional neural network-based automated detection of midline shift in traumatic brain injury cases from head computed tomography scans.

Midline shift (MLS) is a critical indicator of the severity of brain trauma and is even suggestive of changes in intracranial pressure. At present, radiologists have to manually measure the MLS using laborious techniques. Automatic detection of MLS using artificial intelligence can be a cutting-edge solution for emergency health-care personnel to help in prompt diagnosis and treatment. In this study, we sought to determine the accuracy and the prognostic value of our screening tool that automatically detects MLS on computed tomography (CT) images in patients with traumatic brain injuries (TBIs). The study enrolled TBI cases, who presented at the Department of Neurosurgery, All India Institute of Medical Sciences, New Delhi. Institutional ethics committee permission was taken before starting the study. The data collection was carried out for over nine months, i.e., from January 2020 to September 2020. The data collection included head CT scans, patient demographics, clinical details as well as radiologist's reports. The radiologist's reports were considered the "gold standard" for evaluating the MLS. A deep learning-based three dimensional (3D) convolutional neural network (CNN) model was developed using 176 head CT scans. The developed 3D CNN model was trained using 156 scans and was tested on 20 head CTs to determine the accuracy and sensitivity of the model. The screening tool was correctly able to detect 7/10 MLS cases and 4/10 non-MLS cases. The model showed an accuracy of 55% with high specificity (70%) and moderate sensitivity of 40%. An automated solution for screening the MLS can prove useful for neurosurgeons. The results are strong evidence that 3D CNN can assist clinicians in screening MLS cases in an emergency setting.

End-to-end 3D CNN for plot-scale soybean yield prediction using multitemporal UAV-based RGB images

Crop yield prediction from UAV images has significant potential in accelerating and revolutionizing crop breeding pipelines. Although convolutional neural networks (CNN) provide easy, accurate and efficient solutions over traditional machine learning models in computer vision applications, a CNN training requires large number of ground truth data, which is often difficult to collect in the agricultural context. The major objective of this study was to develope an end-to-end 3D CNN model for plot-scale soybean yield prediction using multitemporal UAV-based RGB images with approximately 30,000 sample plots. A low-cost UAV-RGB system was utilized and multitemporal images from 13 different experimental fields were collected at Argentina in 2021. Three commonly used 2D CNN architectures (i.e., VGG, ResNet and DenseNet) were transformed into 3D variants to incorporate the temporal data as the third dimension. Additionally, multiple spatiotemporal resolutions were considered as data input and the CNN architectures were trained with different combinations of input shapes. The results reveal that: (a) DenseNet provided the most efficient result (R2 0.69) in terms of accuracy and model complexity, followed by VGG (R2 0.70) and ResNet (R2 0.65); (b) Finer spatiotemporal resolution did not necessarily improve the model performance but increased the model complexity, while the coarser resolution achieved comparable results; and (c) DenseNet showed lower clustering patterns in its prediction maps compared to the other models. This study clearly identifies that multitemporal observation with UAV-based RGB images provides enough information for the 3D CNN architectures to accurately estimate soybean yield non-destructively and efficiently.

Discriminative analysis of schizophrenia patients using an integrated model combining 3D CNN with 2D CNN: A multimodal MR image and connectomics analysis

ObjectiveFew studies have applied deep learning to the discriminative analysis of schizophrenia (SZ) patients using the fusional features of multimodal MRI data. Here, we proposed an integrated model combining a 3D convolutional neural network (CNN) with a 2D CNN to classify SZ patients. MethodStructural MRI (sMRI) and resting-state functional MRI (rs-fMRI) data were acquired for 140 SZ patients and 205 normal controls. We computed structural connectivity (SC) from the sMRI data as well as functional connectivity (FC), amplitude of low-frequency fluctuation (ALFF), and regional homogeneity (ReHo) from the rs-fMRI data. The 3D images of T1, ReHo, and ALFF were used as the inputs for the 3D CNN model, while the SC and FC matrices were used as the inputs for the 2D CNN model. Moreover, we added squeeze and excitation blocks (SE-blocks) to each layer of the integrated model and used a support vector machine (SVM) to replace the softmax classifier. ResultsThe integrated model proposed in this study, using the fusional features of the T1 images, and the matrices of FC, showed the best performance. The use of the SE-blocks and SVM classifiers significantly improved the performance of the integrated model, in which the accuracy, sensitivity, specificity, area under the curve, and F1-score were 89.86%, 86.21%, 92.50%, 89.35%, and 87.72%, respectively. ConclusionsOur findings indicated that an integrated model combining 3D CNN with 2D CNN is a promising method to improve the classification performance of SZ patients and has potential for the clinical diagnosis of psychiatric diseases.

Distinguishing Laparoscopic Surgery Experts from Novices Using EEG Topographic Features.

The study aimed to differentiate experts from novices in laparoscopic surgery tasks using electroencephalogram (EEG) topographic features. A microstate-based common spatial pattern (CSP) analysis with linear discriminant analysis (LDA) was compared to a topography-preserving convolutional neural network (CNN) approach. Expert surgeons (N = 10) and novice medical residents (N = 13) performed laparoscopic suturing tasks, and EEG data from 8 experts and 13 novices were analysed. Microstate-based CSP with LDA revealed distinct spatial patterns in the frontal and parietal cortices for experts, while novices showed frontal cortex involvement. The 3D CNN model (ESNet) demonstrated a superior classification performance (accuracy > 98%, sensitivity 99.30%, specificity 99.70%, F1 score 98.51%, MCC 97.56%) compared to the microstate based CSP analysis with LDA (accuracy ~90%). Combining spatial and temporal information in the 3D CNN model enhanced classifier accuracy and highlighted the importance of the parietal-temporal-occipital association region in differentiating experts and novices.

Cross-layer connected network with adaptive attention mechanism for 3D multi-organ and tumor segmentations from CT

3D medical image segmentation is one of the most important and challenging tasks for medical image analysis. To enhance the feature extraction ability of 3D convolutional neural networks (CNN) and improve the segmentation accuracy on small tumors, this paper presents a cross-layer connected network with an adaptive attention mechanism to achieve efficient segmentations for 3D multi-organ and tumor from CT. A cross layer connection is proposed to utilize features more effectively by reusing multiple layers of semantic information while recovering spatial information. The pre-activation Residual Block and Squeeze-and-Excitation (SE) block are introduced to enhance the feature extraction ability. Due to the serious class imbalance problem caused by small tumors, the training process of the 3D CNN model is unstable and small targets are difficult to segment. To relieve the training instability problem, Anti-Aliased Pooling (AAP) is introduced to replace the original maxpooling. Moreover, a region of interest (ROI) adaptive attention mechanism based on the auxiliary loss function is developed to improve the segmentation accuracy on small target tumors by combining the target detection method with semantic segmentation models. Extensive experiments show that the dice coefficient of 3D liver tumor segmentation on the LiTS dataset reaches 0.668 by using our designed network, and the experiments on the BRATS dataset for 3D brain tumor segmentation yield average dice coefficient, sensitivity, and Hausdorff distance of 0.706, 0.730, and 5.972 mm, respectively, which are obviously superior to some existing methods.

Wildfire Detection Using Convolutional Neural Networks and PRISMA Hyperspectral Imagery: A Spatial-Spectral Analysis

The exacerbation of wildfires, attributed to the effects of climate change, presents substantial risks to ecological systems, infrastructure, and human well-being. In the context of the Sustainable Development Goals (SDGs), particularly those related to climate action, prioritizing the assessment and management of the occurrence and intensity of extensive wildfires is of utmost importance. In recent times, there has been a significant increase in the frequency and severity of widespread wildfires worldwide, affecting several locations, including Australia, Italy, and the United States of America. The presence of complex phenomena marked by limited predictability leads to significant negative impacts on biodiversity and human lives. The utilization of satellite-derived data with neural networks, such as convolutional neural networks (CNNs), is a potentially advantageous approach for augmenting the monitoring capabilities of wildfires. This research examines the generalization capability of four neural network models, namely the fully connected (FC), one-dimensional (1D) CNN, two-dimensional (2D) CNN, and three-dimensional (3D) CNN model. Each model’s performance, as measured by accuracy, recall, and F1 scores, is assessed through K-fold cross-validation. Subsequently, T-statistics and p-values are computed based on these metrics to conduct a statistical comparison among the different models, allowing us to quantify the degree of similarity or dissimilarity between them. By using training data from Australia and Sicily, the performances of the trained model are evaluated on the test dataset from Oregon. The results are promising, with cross-validation on the training dataset producing mean precision, recall, and F1 scores ranging between approximately 0.97 and 0.98. Especially, the fully connected model has superior generalization capabilities, whilst the 3D CNN offers more refined and less distorted classifications. However, certain issues, such as false fire detection and confusion between smoke and shadows, persist. The aforementioned methodologies offer significant perspectives on the capabilities of neural network technologies in supporting the detection and management of wildfires. These approaches address the crucial matter of domain transferability and the associated dependability of predictions in new regions. This study makes a valuable contribution to the ongoing efforts in climate change by assisting in monitoring and managing wildfires.

Comparative analysis of alignment algorithms for macular optical coherence tomography imaging

BackgroundOptical coherence tomography (OCT) is the most important and commonly utilized imaging modality in ophthalmology and is especially crucial for the diagnosis and management of macular diseases. Each OCT volume is typically only available as a series of cross-sectional images (B-scans) that are accessible through proprietary software programs which accompany the OCT machines. To maximize the potential of OCT imaging for machine learning purposes, each OCT image should be analyzed en bloc as a 3D volume, which requires aligning all the cross-sectional images within a particular volume.MethodsA dataset of OCT B-scans obtained from 48 age-related macular degeneration (AMD) patients and 50 normal controls was used to evaluate five registration algorithms. After alignment of B-scans from each patient, an en face surface map was created to measure the registration quality, based on an automatically generated Laplace difference of the surface map–the smoother the surface map, the smaller the average Laplace difference. To demonstrate the usefulness of B-scan alignment, we trained a 3D convolutional neural network (CNN) to detect age-related macular degeneration (AMD) on OCT images and compared the performance of the model with and without B-scan alignment.ResultsThe mean Laplace difference of the surface map before registration was 27 ± 4.2 pixels for the AMD group and 26.6 ± 4 pixels for the control group. After alignment, the smoothness of the surface map was improved, with a mean Laplace difference of 5.5 ± 2.7 pixels for Advanced Normalization Tools Symmetric image Normalization (ANTs-SyN) registration algorithm in the AMD group and a mean Laplace difference of 4.3 ± 1.4.2 pixels for ANTs in the control group. Our 3D CNN achieved superior performance in detecting AMD, when aligned OCT B-scans were used (AUC 0.95 aligned vs. 0.89 unaligned).ConclusionsWe introduced a novel metric to quantify OCT B-scan alignment and compared the effectiveness of five alignment algorithms. We confirmed that alignment could be improved in a statistically significant manner with readily available alignment algorithms that are available to the public, and the ANTs algorithm provided the most robust performance overall. We further demonstrated that alignment of OCT B-scans will likely be useful for training 3D CNN models.

Predicting the efficacy of non-steroidal anti-inflammatory drugs in migraine using deep learning and three-dimensional T1-weighted images

Deep learning (DL) models based on individual images could contribute to tailored therapies and personalized treatment strategies. We aimed to construct a DL model using individual 3D structural images for predicting the efficacy of non-steroidal anti-inflammatory drugs (NSAIDs) in migraine. A 3D convolutional neural network model was constructed, with ResNet18 as the classification backbone, to link structural images to predict the efficacy of NSAIDs. In total, 111 patients were included and allocated to the training and testing sets in a 4:1 ratio. The prediction accuracies of the ResNet34, ResNet50, ResNeXt50, DenseNet121, and 3D ResNet18 models were 0.65, 0.74, 0.65, 0.70, and 0.78, respectively. This model, based on individual 3D structural images, demonstrated better predictive performance in comparison to conventional models. Our study highlights the feasibility of the DL algorithm based on brain structural images and suggests that it can be applied to predict the efficacy of NSAIDs in migraine treatment.

Automated intracranial hemorrhage detection in traumatic brain injury using 3D CNN.

Intracranial hemorrhage (ICH) is a prevalent and potentially fatal consequence of traumatic brain injury (TBI). Timely identification of ICH is crucial to ensure timely intervention and to optimize better patient outcomes. However, the current methods for diagnosing ICH from head computed tomography (CT) scans require skilled personnel (Radiologists and/or Neurosurgeons) who may be unavailable in all centers, especially in rural areas. The aim of this study is to develop a neurotrauma screening tool for identifying ICH from head CT scans of TBI patients. We prospectively collected head CT scans from the Department of Neurosurgery, All India Institute of Medical Sciences, New Delhi. Approximately 738 consecutive head CT scans from patients enrolled in the department were collected for this study spanning a duration of 9 months, that is, January 2020 to September 2020. The metadata collected along with the head CT scans consisted of demographic and clinical details and the radiologist's report which was used as the gold standard. A deep learning-based 3D convolutional neural network (CNN) model was trained on the dataset. The pre-processing, hyperparameters, and augmentation were common for training the 3D CNN model whereas the training modules were set differently. The model was trained along with the save best model option and was monitored by validation metrics. The Institute Ethics Committee permission was taken before starting the study. We developed a 3D CNN model for automatically detecting the ICH from head CT scans. The screening tool was tested in 20 cases and trained on 200 head CT scans, with 99 normal head CT and 101 CT scans with some type of ICH. The final model performed with 90% sensitivity, 70% specificity, and 80% accuracy. Our study reveals that the automated screening tool exhibits a commendable level of accuracy and sensitivity in detecting ICH from the head CT scans. The results indicate that the 3D CNN approach has a potential for further exploring the TBI-related pathologies.

Metallogenic Prediction of the Zaozigou Gold Deposit Using 3D Geological and Geochemical Modeling

Deep-seated mineralization prediction is an important scientific problem in the area of mineral resources exploration. The 3D metallogenic information extraction of geology and geochemistry can be of great help. This study uses 3D modeling technology to intuitively depict the spatial distribution of orebodies, fractures, and intrusive rocks. In particular, the geochemical models of 12 elements are established for geochemical metallogenic information extraction. Subsequently, the front halo element association of As-Sb-Hg, the near-ore halo element association of Au-Ag-Cu-Pb-Zn, and the tail halo element association of W-Mo-Bi are identified. Upon this foundation, the 3D convolutional neural network model is built and used for deep-seated mineralization prediction, which expresses a high performance (AUC = 0.99). Associated with the metallogenic regularity, two mineral exploration targets are delineated, which might be able to serve as beneficial achievements for deep exploration in the Zaozigou gold deposit.

3DCNN predicting brain age using diffusion tensor imaging.

Neuroimaging-based brain age prediction using deep learning is gaining popularity. However, few studies have attempted to leverage diffusion tensor imaging (DTI) to predict brain age. In this study, we proposed a 3D convolutional neural network model (3DCNN) and trained it on fractional anisotropy (FA) data from six publicly available datasets (n = 2406, age = 17-60) to estimate brain age. Implementing a two-loop nested cross-validation scheme with a tenfold cross-validation procedure, we achieved a robust prediction performance of a mean absolute error (MAE) of 2.785 and a correlation coefficient of 0.932. We also employed Grad-Cam++ to visualize the salient features of the proposed model. We identified a few highly salient fiber tracts, including the genu of corpus callosum and the left cerebellar peduncle, among others that play a pivotal role in our model. In sum, our model reliably predicted brain age and provided novel insight into age-related changes in brains' axonal structure.

An uncertainty-aware self-training framework with consistency regularization for the multilabel classification of common computed tomography signs in lung nodules.

Computed tomography (CT) signs of lung nodules play an important role in indicating lung nodules' malignancy, and accurate automatic classification of these signs can help doctors improve their diagnostic efficiency. However, few relevant studies targeting multilabel classification (MLC) of nodule signs have been conducted. Moreover, difficulty in obtaining labeled data also restricts this avenue of research to a large extent. To address these problems, a multilabel automatic classification system for nodule signs is proposed, which consists of a 3-dimensional (3D) convolutional neural network (CNN) and an efficient new semi-supervised learning (SSL) framework. Two datasets were used in our experiments: Lung Nodule Analysis 16 (LUNA16), a public dataset for lung nodule classification, and a private dataset of nodule signs. The private dataset contains 641 nodules, 454 of which were annotated with 6 important signs by radiologists. Our classification system consists of 2 main parts: a 3D CNN model and an SSL method called uncertainty-aware self-training framework with consistency regularization (USC). In the system, supervised training is performed with labeled data, and simultaneously, an uncertainty-and-confidence-based strategy is used to select pseudo-labeled samples for unsupervised training, thus jointly realizing the optimization of the model. For the MLC of nodule signs, our proposed 3D CNN achieved satisfactory results with a mean average precision (mAP) of 0.870 and a mean area under the curve (AUC) of 0.782. In semi-supervised experiments, compared with supervised learning, our proposed SSL method could increase the mAP by 7.6% (from 0.730 to 0.806) and the mean AUC by 8.1% (from 0.631 to 0.712); it thus efficiently utilized the unlabeled data and achieved superior performance improvement compared to the recently advanced methods. We realized the optimal MLC of lung nodule signs with our proposed 3D CNN. Our proposed SSL method can also offer an efficient solution for the insufficiency of labeled data that may exist in the MLC tasks of 3D medical images.

Video Vision Transformer를 이용한 벼의 수확량 예측과 Self-attention 시각화

The government and farmers' organizations are paying much attention to the problem of predicting how much rice can be produced each year. However, it is difficult to accurately predict the yield of rice due to variable factors such as extreme climate change and various pests and diseases that change every year. In this study, images were collected several times during the growing season of rice through a multi-spectral sensor mounted on an unmanned aerial vehicle, and rice yield was predicted using a deep learning algorithm. Multispectral images can be viewed as a kind of image data taken several times at regular intervals, and rice yield was predicted using the Video Vision Transformer (ViViT) model, which applies the Transformer structure to image computer vision among deep learning algorithms. The ViViT model generates patches by dividing the input image into a certain size, and as a result of learning the model by setting the size of these patches differently, it was found that the smaller the patch size, the better the predictive power. In addition, as a result of comparing prediction performance with a 3D CNN model that receives an image as an input in a CNN (Convolutional Neural Network) structure used in the image processing field, it was found that the ViViT model using a small patch size performed better. As a result of visualizing the weight matrix of the ViViT model as a heat map, images taken in mid- to late August appear to be important in yield prediction, making it possible to predict yield about two months before rice harvest.

Classification of Pulmonary Nodules in 2-[18F]FDG PET/CT Images with a 3D Convolutional Neural Network

Purpose2-[18F]FDG PET/CT plays an important role in the management of pulmonary nodules. Convolutional neural networks (CNNs) automatically learn features from images and have the potential to improve the discrimination between malignant and benign pulmonary nodules. The purpose of this study was to develop and validate a CNN model for classification of pulmonary nodules from 2-[18F]FDG PET images.MethodsOne hundred thirteen participants were retrospectively selected. One nodule per participant. The 2-[18F]FDG PET images were preprocessed and annotated with the reference standard. The deep learning experiment entailed random data splitting in five sets. A test set was held out for evaluation of the final model. Four-fold cross-validation was performed from the remaining sets for training and evaluating a set of candidate models and for selecting the final model. Models of three types of 3D CNNs architectures were trained from random weight initialization (Stacked 3D CNN, VGG-like and Inception-v2-like models) both in original and augmented datasets. Transfer learning, from ImageNet with ResNet-50, was also used.ResultsThe final model (Stacked 3D CNN model) obtained an area under the ROC curve of 0.8385 (95% CI: 0.6455–1.0000) in the test set. The model had a sensibility of 80.00%, a specificity of 69.23% and an accuracy of 73.91%, in the test set, for an optimised decision threshold that assigns a higher cost to false negatives.ConclusionA 3D CNN model was effective at distinguishing benign from malignant pulmonary nodules in 2-[18F]FDG PET images.

Predicting gas diffusion layer flow information in proton exchange membrane fuel cells from cross-sectional data using deep learning methods

Obtaining transient flow field information of gas diffusion layers (GDLs) is a crucial issue for improving the performance of proton exchange membrane fuel cells (PEMFCs). While physically-based methods, such as effective medium theory and solving partial differential equations, can be utilized to calculate fluid flow in GDL, the computational cost associated with such methods remains substantial. In this study, a 3D multiphase flow dataset of GDLs is obtained using fluid of volume (VOF) calculation, and the 3D data is sliced. The resulting cross-sectional data are used to establish a convolutional neural network (CNN) model based on two-dimensional (2D) data for predicting fluid flow in GDLs. The reliability of predicting the saturation information of 3D GDLs using 2D cross-sectional data is verified and discussed. Subsequently, a comparison is made between the accuracy and computational cost of predicting the saturation of 3D GDLs using 2D and 3D CNN models. The results indicate that using 25 cross-sectional images provide accurate predictions of GDL saturation in 3D. When using cross-sectional images as inputs, 100 × 100 images are found to be more representative than 50 × 50 images, resulting in an average increase of 1.86% and 13.36% in R2 and RMSE, respectively. While the 3D CNN model outperforms the 2D CNN model in predicting GDL saturation in 3D by only 0.62% in terms of R2 score, its computational cost is two orders of magnitude higher. These findings suggest that 2D GDL cross-sectional images can also be used for predicting 3D GDL flow information and effectively reducing computational costs. The findings of this study provide a profound insight into the GDL flow phenomena and contribute to the development of more efficient and accurate fuel cell models.