Hierarchical Feature Extraction Research Articles

Detection and Classification of Banana Leaf Diseases: Systematic Literature Review

Bananas, a staple fruit globally, are essential for sustenance, employment, and income. However, diseases like Sigatoka, Bacterial Wilt, Bunchy Top, and Fusarium Wilt pose a threat to their cultivation, affecting both small-scale and large-scale production. This survey investigates methods for the early identification and classification of these banana leaf diseases using deep learning and machine learning techniques. A systematic review of 15 studies revealed that the majority of research concentrates on binary classification, which distinguishes healthy from diseased leaves. Common preprocessing steps include image resizing, color space conversion, and background removal to improve model accuracy. We utilize techniques such as ensemble approaches, support vector machines (SVM), random forests, K-means clustering, and convolutional neural networks (CNNs), with CNNs demonstrating superior performance, achieving accuracy rates ranging from 85% to 98.97%. CNNs excel in hierarchical feature extraction but require significant computational power. Traditional machine learning methods offer simplicity and resistance to overfitting but need careful parameter tuning. Advanced deep learning architectures, such as DenseNet and Inception V3, achieve high accuracy but with greater computational demands. Lightweight models like SqueezeNet balance performance and size, but ensemble methods, while improving generalization, add complexity. The choice of method depends on dataset characteristics, available computational resources, and desired trade-offs between performance and complexity. This study provides an overview of current research in banana leaf disease classification, discussing the strengths and limitations of various approaches and suggesting directions for future research to improve detection accuracy and robustness.

Machine learning based battery pack health prediction using real-world data

The complex operational conditions in real-world electric vehicles (EVs) contribute to the complexity of managing and maintaining battery packs. Adding to these challenges is the intricate task of modeling the inconsistent coupling among individual cells within these packs. This study addresses the ongoing challenges in modeling lithium-ion battery (LIB) cells within packs and estimating their state of health (SOH) for practical applications. This research proposed a PCA-CNN-Transformer method to model and predict the SOH model of real-world EV. Three main contributions are presented: a novel approach to defining an attenuation SOH model based on delivered energy, a methodology utilizing Principal Component Analysis (PCA) for cell modeling, and an SOH estimation model employing CNN-Transformer architecture. To address both pack and cell-level modeling, a hierarchical feature extraction approach is proposed. The health features extracted from both levels are assessed using grey relational analysis, showing a strong correlation with LIB SOH, exceeding 0.70. The proposed cell modeling method significantly reduces data size by 96%, enhancing computational efficiency. Furthermore, the integration of 1D-CNN in the SOH estimation model overcomes the limitations of the attention mechanism, achieving a MAE with 0.0406 and r-square of 0.9327, improved the original transformer network performance by 10.95%. This study also examines and discusses the performance of the informer and transformer models, elaborating why the informer model underperformed in this dataset.

NCNet: Deformable medical image registration network based on neighborhood cross-attention combined with multi-resolution constraints

Objective. Existing registration networks based on cross-attention design usually divide the image pairs to be registered into patches for input. The division and merging operations of a series of patches are difficult to maintain the topology of the deformation field and reduce the interpretability of the network. Therefore, our goal is to develop a new network architecture based on a cross-attention mechanism combined with a multi-resolution strategy to improve the accuracy and interpretability of medical image registration. Approach. We propose a new deformable image registration network NCNet based on neighborhood cross-attention combined with multi-resolution strategy. The network structure mainly consists of a multi-resolution feature encoder, a multi-head neighborhood cross-attention module and a registration decoder. The hierarchical feature extraction capability of our encoder is improved by introducing large kernel parallel convolution blocks; the cross-attention module based on neighborhood calculation is used to reduce the impact on the topology of the deformation field and double normalization is used to reduce its computational complexity. Main result. We performed atlas-based registration and inter-subject registration tasks on the public 3D brain magnetic resonance imaging datasets LPBA40 and IXI respectively. Compared with the popular VoxelMorph method, our method improves the average DSC value by 7.9% and 3.6% on LPBA40 and IXI. Compared with the popular TransMorph method, our method improves the average DSC value by 4.9% and 1.3% on LPBA40 and IXI. Significance. We proved the advantages of the neighborhood attention calculation method compared to the window attention calculation method based on partitioning patches, and analyzed the impact of the pyramid feature encoder and double normalization on network performance. This has made a valuable contribution to promoting the further development of medical image registration methods.

NHi-SEGA: n-Hierarchy SEmantic Guided Attention for few-shot learning

AbstractHumans excel at learning and recognizing objects, swiftly adapting to new concepts with just a few samples. However, current studies in computer vision on few-shot learning have not yet achieved human performance in integrating prior knowledge during the learning process. Humans utilize a hierarchical structure of object categories based on past experiences to facilitate learning and classification. Therefore, we propose a method named n-Hierarchy SEmantic Guided Attention (nHi-SEGA) that acquires abstract superclasses. This allows the model to associate with and pay attention to different levels of objects utilizing semantics and visual features embedded in the class hierarchy (e.g., house finch-bird-animal, goldfish-fish-animal, rose-flower-plant), resembling human cognition. We constructed an nHi-Tree using WordNet and Glove tools and devised two methods to extract hierarchical semantic features, which were then fused with visual features to improve sample feature prototypes.

Intelligent inversion of mesoscale eddy temperature anomaly profiles based on multi-source remote sensing data

Inversion of the underwater temperature anomaly of mesoscale eddies from sea surface remote sensing data plays an important role in understanding the three-dimensional structure of eddies. In this article, a neural network structure, named the eddy temperature anomaly inversion network (EddyTAINet), is proposed and established to invert the vertical temperature anomalies of eddies using 20-year (2002–2022) Argo profiles and multi-source remote sensing data in the northwestern Pacific Ocean. Vertical temperature anomaly profiles with 45 depth levels ranging from 10 to 1000 m are derived as the outputs of the inversion networks. The input feature sequences are extracted from the mesoscale eddy product based on sea level anomaly (SLA) data, and sea surface temperature (SST) data are also introduced into the proposed network structure to improve its accuracy. For the problem of inconsistent sizes of the 2D matrix SST data corresponding to mesoscale eddies, the spatial pyramid pooling (SPP) network module is introduced, which allows different sizes of SST data to be simply convolved and then subjected to hierarchical feature extraction to obtain fixed-size feature vectors. Validation and comparison of the network structure are performed and show the improved performance of EddyTAINet. The average root mean square error (RMSE) and mean absolute error (MAE) are 0.63 °C/0.66 °C and 0.48 °C/0.51 °C for cyclonic and anticyclonic eddies, respectively, which are more than 20% lower than the results of the network using the SLA-based data.

Research On Cultivating the Employment Concept of College Students with The Great Spirit of Building the Party - Based on A Data-Driven Approach

One of the most essential ways to see the worth of college students is via employment, which is also the main focus of student development programs for the Spirit of Building the Party. Some real-world issues with the conventional approach to college career counselling stem from the disconnection between students' education and the real world. For example, there is often a misalignment between students' social needs and their comprehensive skill sets and between students' job-seeking needs and their actual employment situations. There has to be extensive research into how to better assist college graduates in finding work that suits their skills and experience so that society can fulfil the practical demands of its growth in this area for the Spirit of Building the Party. Problems in the workplace have garnered a lot of attention because they threaten society's steady progress and the quality of life for its citizens. The nation now needs to tackle the issue of challenging employment for the Spirit of Building the Party. Institutions of higher learning have responded to societal and economic demands by emphasizing the development of students' social abilities and increasing their access to the job market. To provide a solid groundwork for creating employment assistance programs, this research suggests an analytical strategy for college students' jobs based on a data-driven approach for the Spirit of Building the Party. The proposed model is used to plan the supply and structure of skills from a talent supply standpoint. In contrast, the regression model forecasts university students' employment requirements for the Spirit of Building the Party. The data mining technique used in this work combines enhanced fuzzy hierarchical clustering and feature extraction based on semantic similarity correlation. The testing findings showed that the method performed better in categorizing data and could easily handle and analyze big-text datasets

Learning feature relationships in CNN model via relational embedding convolution layer

Establishing the relationships among hierarchical visual attributes of objects in the visual world is crucial for human cognition. The classic convolution neural network (CNN) can successfully extract hierarchical features but ignore the relationships among features, resulting in shortcomings compared to humans in areas like interpretability and domain generalization. Recently, algorithms have introduced feature relationships by external prior knowledge and special auxiliary modules, which have been proven to bring multiple improvements in many computer vision tasks. However, prior knowledge is often difficult to obtain, and auxiliary modules bring additional consumption of computing and storage resources, which limits the flexibility and practicality of the algorithm. In this paper, we aim to drive the CNN model to learn the relationships among hierarchical deep features without prior knowledge and consumption increasing, while enhancing the fundamental performance of some aspects. Firstly, the task of learning the relationships among hierarchical features in CNN is defined and three key problems related to this task are pointed out, including the quantitative metric of connection intensity, the threshold of useless connections, and the updating strategy of relation graph. Secondly, Relational Embedding Convolution (RE-Conv) layer is proposed for the representation of feature relationships in convolution layer, followed by a scheme called use & disuse strategy which aims to address the three problems of feature relation learning. Finally, the improvements brought by the proposed feature relation learning scheme have been demonstrated through numerous experiments, including interpretability, domain generalization, noise robustness, and inference efficiency. In particular, the proposed scheme outperforms many state-of-the-art methods in the domain generalization community and can be seamlessly integrated with existing methods for further improvement. Meanwhile, it maintains comparable precision to the original CNN model while reducing floating point operations (FLOPs) by approximately 50%.

Short-term wind power forecasting based on multi-scale receptive field-mixer and conditional mixture copula

Integrating renewable wind power (WP) into the grid exacerbates variability and challenges reliability. Establishing an effective forecasting system is crucial for risk avoidance, but achieving effective, interpretable, and accurate predictions remains an obstacle due to the high volatility and uncertainty of WP. This study proposes an innovative and trustworthy wind power forecasting system combined with point and interval predictions. Specifically, the adaptive multi-scale convolutional receptive field (AMSCF) combined with gated recurrent unit (GRU) for WP point prediction is proposed to capture the spatiotemporal dependencies of WP generation. This AMSCF method uses different size receptive fields, fuses them using channel dimension concatenation, operates pooling layer and adaptive fusion through field attention, which extracts hierarchical features for WP forecasting. Then conditional mixture Copula (Con-mCopula) function integrated with multiobjective optimizer is established for WP interval prediction, which reduces the WP interval prediction accuracy error as some information neglected. The wind power cluster dataset from Germany was selected for experiments. Based on the results, for both datasets, the designed prediction system can achieve better point and interval prediction performance resulting in a reduction of mean absolute error (MAE) by approximately 40 % compared to other existing components in the market. This study deconstructs the “Black Box” in deep network for energy forecasting through interpretability analysis, which accelerates the development of AI technology in WP forecasting system and provides decision-makers with trustworthy WP forecasting assistance.

Early detection of autism spectrum disorder via deep-learning application of fMRI and machine learning for ASD children identifications

This study unveils an advanced convolutional-neural-network (CNN) algorithm that was meticulously engineered to examine resting-state functional magnetic resonance imaging (fMRI) for early ASD detection in pediatric cohorts. The CNN architecture amalgamates convolutional, pooling, batch-normalization, dropout, and fully connected layers, optimized for high-dimensional data interpretation. Rigorous preprocessing yielded 22,176 two-dimensional echo planar samples from 126 subjects (56 ASD, 70 controls) who were sourced from the Autism Brain Imaging Data Exchange (ABIDE I) repository. The model, trained on 17,740 samples across 50 epochs, demonstrated unparalleled diagnostic metrics – accuracy of 99.39%, recall of 98.80%, precision of 99.85%, and an F1 score of 99.32% – and thereby eclipsed extant computational methodologies. Feature map analyses substantiated the model’s hierarchical feature extraction capabilities. This research elucidates a deep learning framework for computer-assisted ASD screening via fMRI, with transformative implications for early diagnosis and intervention. And, this study addresses the critical need for early detection and intervention in autism spectrum disorder (ASD) using machine learning. Specific therapies are needed for ASD, a neurodevelopmental disease that affects social interaction and communication. To find trends in ASD, our research uses a variety of early childhood screening tests as training sets for machine learning algorithms. The methodology that has been suggested utilizes methods of machine learning to compute the ASD spectrum, considering its many expressions. By using multidisciplinary methods and sophisticated screening instruments, we want to create an accurate system for early ASD detection. Algorithmic transparency, data protection, and ethical considerations are essential. This study seeks to build precise instruments for early ASD detection by promoting collaboration between specialists in neurodevelopment, psychology, and machine learning. A robust instrument that enhances the knowledge of medical practitioners is machine learning. Results show how innovation may transform early interventions and help people on the autistic spectrum achieve enhanced results.

HiRENet: Novel convolutional neural network architecture using Hilbert-transformed and raw electroencephalogram (EEG) for subject-independent emotion classification

Background and objectivesConvolutional neural networks (CNNs) are the most widely used deep-learning framework for decoding electroencephalograms (EEGs) due to their exceptional ability to extract hierarchical features from high-dimensional EEG data. Traditionally, CNNs have primarily utilized multi-channel raw EEG data as the input tensor; however, the performance of CNN-based EEG decoding may be enhanced by incorporating phase information alongside amplitude information. MethodsThis study introduces a novel CNN architecture called the Hilbert-transformed (HT) and raw EEG network (HiRENet), which incorporates both raw and HT EEG as inputs. This concurrent use of HT and raw EEG aims to integrate phase information with existing amplitude information, potentially offering a more comprehensive reflection of functional connectivity across various brain regions. The HiRENet model was developed using two CNN frameworks: ShallowFBCSPNet and a CNN with a residual block (ResCNN). The performance of the HiRENet model was assessed using a lab-made EEG database to classify human emotions, comparing three input modalities: raw EEG, HT EEG, and a combination of both signals. Additionally, the computational complexity was evaluated to validate the computational efficiency of the ResCNN design. ResultsThe HiRENet model based on ResCNN achieved the highest classification accuracy, with 86.03% for valence and 84.01% for arousal classifications, surpassing traditional CNN methodologies. Considering computational efficiency, ResCNN demonstrated superiority over ShallowFBCSPNet in terms of speed and inference time, despite having a higher parameter count. ConclusionOur experimental results showed that the proposed HiRENet can be potentially used as a new option to improve the overall performance for deep learning-based EEG decoding problems.

AdaTreeFormer: Few shot domain adaptation for tree counting from a single high-resolution image

The process of estimating and counting tree density using only a single aerial or satellite image is a difficult task in the fields of photogrammetry and remote sensing. However, it plays a crucial role in the management of forests. The huge variety of trees in varied topography severely hinders tree counting models to perform well. The purpose of this paper is to propose a framework that is learnt from the source domain with sufficient labeled trees and is adapted to the target domain with only a limited number of labeled trees. Our method, termed as AdaTreeFormer, contains one shared encoder with a hierarchical feature extraction scheme to extract robust features from the source and target domains. It also consists of three subnets: two for extracting self-domain attention maps from source and target domains respectively and one for extracting cross-domain attention maps. For the latter, an attention-to-adapt mechanism is introduced to distill relevant information from different domains while generating tree density maps; a hierarchical cross-domain feature alignment scheme is proposed that progressively aligns the features from the source and target domains. We also adopt adversarial learning into the framework to further reduce the gap between source and target domains. Our AdaTreeFormer is evaluated on six designed domain adaptation tasks using three tree counting datasets, i.e. Jiangsu, Yosemite, and London. Experimental results show that AdaTreeFormer significantly surpasses the state of the art, e.g. in the cross domain from the Yosemite to Jiangsu dataset, it achieves a reduction of 15.9 points in terms of the absolute counting errors and an increase of 10.8% in the accuracy of the detected trees’ locations. The codes and datasets are available at https://github.com/HAAClassic/AdaTreeFormer.

Intelligent Advanced Attack Detection Technology based on Multi-modal Data Fusion

This paper proposes an adaptive Wedman intrusion detection algorithm (AID-DFS) for data fusion. Firstly, feature extraction of abnormal text detection is carried out using a BI-gated loop (Bi-GRU). Multi-branch convolutional recurrent neural network (CNN-RNN) extracts hierarchical features from abnormal images. The multi-mode dynamic fusion uses the intermodal and intramodal attention mechanisms. In this way, a joint representation of multiple modes is obtained. The visual perception mechanism is used to realize multichannel integration and strengthen the function of original information in multichannel. The experimental results show that the proposed method has 99.6% accuracy and 94.9% accuracy. Compared with other algorithms, the proposed method can improve the performance of the intrusion system by about 10.2%.

Food Image Classification and Calorie Prediction for Dietary Analysis

Abstract: This paper presents a unique method to classify food items and estimate the calorie content based on photos by combining Convolutional Neural Networks (CNNs) with image processing techniques. To improve the quality of the dataset, we first curate a wide range of food photographs from different presentation styles and cuisines. We do this by applying preprocessing techniques including image segmentation and feature extraction. Next, using the CNN's capacity to extract hierarchical features from raw pixel data, a custom deep CNN architecture is trained on this dataset to efficiently classify diverse food items and achieve high accuracy in differentiating between different dishes. Furthermore, we tackle the problem of calorie prediction by employing regression models that include features extracted from the CNN, which allows one to forecast the calorie content of a food item based on its visual attributes. Our technique intends to enable users to make more educated nutritional decisions and better control their caloric intake by fusing image categorization with calorie prediction. The suggested approach shows encouraging results in terms of food item classification accuracy and precision in calorie estimate. Its possible uses include promoting better eating practices and assisting individuals, dietitians, and the food industry with their dietary monitoring. In addition, our technology gives customers access to weekly calorie intake information, which can help them avoid obesity-related illnesses like diabetes.

Hierarchical Predictions of Fine-to-Coarse Time Span and Atmospheric Field Reconstruction for Typhoon Track Prediction

The prediction of typhoon tracks in the Northwest Pacific is key to reducing human casualties and property damage. Traditional numerical forecasting models often require substantial computational resources, are high-cost, and have significant limitations in prediction speed. This research is dedicated to using deep learning methods to address the shortcomings of traditional methods. Our method (AFR-SimVP) is based on a large-kernel convolutional spatio-temporal prediction network combined with multi-feature fusion for forecasting typhoon tracks in the Northwest Pacific. In order to more effectively suppress the effect of noise in the dataset to enhance the generalization ability of the model, we use a multi-branch structure, incorporate an atmospheric reconstruction subtask, and propose a second-order smoothing loss to further improve the prediction ability of the model. More importantly, we innovatively propose a multi-time-step typhoon prediction network (HTAFR-SimVP) that does not use the traditional recurrent neural network family of models at all. Instead, through fine-to-coarse hierarchical temporal feature extraction and dynamic self-distillation, multi-time-step prediction is achieved using only a single regression network. In addition, combined with atmospheric field reconstruction, the network achieves integrated prediction for multiple tasks, which greatly enhances the model’s range of applications. Experiments show that our proposed network achieves optimal performance in the 24 h typhoon track prediction task. Our regression network outperforms previous recurrent network-based typhoon prediction models in the multi-time-step prediction task and also performs well in multiple integration tasks.

Heritage Identification of Monument Using Deep Learning

Abstract: There are many beautiful and colorful countries in the world which are famous for their culture and heritage. The historical landmarks of such countries convey a great deal through their magnificent structures and rich heritage. The extensive past and legacy of such nations are apparent in the splendid palaces, fortresses, towers, shrines, and cathedrals that are scattered throughout the entirety of their land. We need to protect cultural monuments and landmarks as they are one of the sources of inspiration for future generations. Traditional methods for monument identification often rely on manual inspection or rudimentary computer vision techniques, which can be time-consuming and error-prone. We can provide the best solutions using modern machine learning and deep learning techniques. In our model, we used an approach based on deep learning for the automated identification of heritage monuments from images. We leveraged the power of convolutional neural networks (CNNs) to extract hierarchical features from monument images and train a classifier to recognize different architectural styles and structures. We evaluated our approach on a dataset comprising diverse monument images collected from public archives and online repositories. We applied various pre-processing techniques such as image resizing, normalization of pixel values, and data augmentation. We analyzed and further improved the model’s performance for heritage identification of monuments.

SwinSOD: Salient object detection using swin-transformer

The Transformer structure has achieved excellent performance in a wide range of applications in computer vision, and Swin-Transformer also shows strong feature representation capabilities. On this basis, we proposed a fusion model SwinSOD for RGB salient object detection. This model used a Swin-Transformer as the encoder to extract hierarchical features, was driven by a multi-head attention mechanism to bridge the gap between hierarchical features, progressively fused adjacent layer feature information under the guidance of global information, and refined the boundaries of saliency objects through the feedback information. Specifically, the Swin-Transformer encoder extracted multi-level features and then recalibrated the channels to optimize intra-layer channel features. The feature fusion module realized feature fusion between each layer under the guidance of global information. In order to clarify the fuzzy boundaries, the second stage feature fusion achieved edge refinement under the guidance of feedback information. The proposed model outperforms state-of-the-art models on five popular SOD datasets, demonstrating the advanced performance of this network. Code released: https://github.com/user-wu/SwinSOD.

Research of news text classification method based on hierarchical semantics and prior correction

News text is an important branch of natural language processing. Compared to ordinary texts, news text has significant economic and scientific value. The characteristics of news text include structural hierarchy, diverse label categories, and limited high-quality annotation samples. Many machine learning and deep learning methods exist to analyze various forms of news text. However, due to label imbalance, hierarchical semantics, and confusing labels, current methods have limitations. Therefore, this paper proposes a news text classification framework based on hierarchical semantics and prior correction (HSPC). Firstly, data augmentation is used to enhance the diversity of the training set and adversarial learning is employed to improve the resistance of the model with its robustness. Then, a hierarchical feature extraction approach is employed to extract semantic features from different levels of news texts. Consequentially, a feature fusion method is designed to allow the model to focus on relevant hierarchical semantics for label classification. Finally, highly confusing label predictions are corrected to optimize the label prediction of the model and improve confidence. Multiple experiments are performed on four widely used public datasets. The experimental results indicate that HSPC achieves higher classification accuracy compared to other models. On the FCT, AGNews, THUCNews, and Ohsumed datasets, HSPC improves the accuracy by 1.03%, 1.38%, 2.55%, and 1.15%, respectively, compared to state-of-the-art methods. This validates the rationality and effectiveness of the designed mechanisms.

Histopathology language-image representation learning for fine-grained digital pathology cross-modal retrieval

Large-scale digital whole slide image (WSI) datasets analysis have gained significant attention in computer-aided cancer diagnosis. Content-based histopathological image retrieval (CBHIR) is a technique that searches a large database for data samples matching input objects in both details and semantics, offering relevant diagnostic information to pathologists. However, the current methods are limited by the difficulty of gigapixels, the variable size of WSIs, and the dependence on manual annotations. In this work, we propose a novel histopathology language-image representation learning framework for fine-grained digital pathology cross-modal retrieval, which utilizes paired diagnosis reports to learn fine-grained semantics from the WSI. An anchor-based WSI encoder is built to extract hierarchical region features and a prompt-based text encoder is introduced to learn fine-grained semantics from the diagnosis reports. The proposed framework is trained with a multivariate cross-modal loss function to learn semantic information from the diagnosis report at both the instance level and region level. After training, it can perform four types of retrieval tasks based on the multi-modal database to support diagnostic requirements. We conducted experiments on an in-house dataset and a public dataset to evaluate the proposed method. Extensive experiments have demonstrated the effectiveness of the proposed method and its advantages to the present histopathology retrieval methods. The code is available at https://github.com/hudingyi/FGCR.

SAH-NET: Structure-Aware Hierarchical Network for Clustered Microcalcification Classification in Digital Breast Tomosynthesis.

Benign and malignant classification of clustered microcalcifications (MCs) in digital breast tomosynthesis (DBT) is an essential task in computer-aided diagnosis. However, due to the anisotropic resolution of DBT, three-dimensional (3-D) convolutional neural network (CNN)-based methods cannot extract hierarchical features efficiently. Moreover, the sparse distribution of MC points in the cluster makes it difficult for the CNN to extract discriminative structural information for classification. To comprehensively address these challenges, we propose a novel structure-aware hierarchical network (SAH-Net) for benign and malignant classification of clustered MC in a DBT volume. Specifically, the two-dimensional (2-D) group convolution is used to extract intraslice features. The one-to-one correspondence between group convolutions and slices ensures the independence of hierarchical feature extraction. Then, a partial deformable Transformer-based 3-D structural feature learning module is proposed to capture the long-range dependency between MC points in the cluster. We evaluate the proposed method on an in-house dataset with 495 clustered MCs collected from 462 DBT images. Experimental results confirm the validity of our proposed modules. The results also show that the proposed SAH-Net outperforms several other representative methods on this topic, and achieves the best classification result, with an area under the receiver operation curve (AUC) of 86.87%. The implementation of the proposed model is available at https://github.com/sunhaotian130911/SAHNet.

Susceptibility-Weighted MRI for Predicting NF-2 Mutations and S100 Protein Expression in Meningiomas.

S100 protein expression levels and neurofibromatosis type 2 (NF-2) mutations result in different disease courses in meningiomas. This study aimed to investigate non-invasive biomarkers of NF-2 copy number loss and S100 protein expression in meningiomas using morphological, radiomics, and deep learning-based features of susceptibility-weighted MRI (SWI). This retrospective study included 99 patients with S100 protein expression data and 92 patients with NF-2 copy number loss information. Preoperative cranial MRI was conducted using a 3T clinical MR scanner. Tumor volumes were segmented on fluid-attenuated inversion recovery (FLAIR) and subsequent registration of FLAIR to high-resolution SWI was performed. First-order textural features of SWI were extracted and assessed using Pyradiomics. Morphological features, including the tumor growth pattern, peritumoral edema, sinus invasion, hyperostosis, bone destruction, and intratumoral calcification, were semi-quantitatively assessed. Mann-Whitney U tests were utilized to assess the differences in the SWI features of meningiomas with and without S100 protein expression or NF-2 copy number loss. A logistic regression analysis was used to examine the relationship between these features and the respective subgroups. Additionally, a convolutional neural network (CNN) was used to extract hierarchical features of SWI, which were subsequently employed in a light gradient boosting machine classifier to predict the NF-2 copy number loss and S100 protein expression. NF-2 copy number loss was associated with a higher risk of developing high-grade tumors. Additionally, elevated signal intensity and a decrease in entropy within the tumoral region on SWI were observed in meningiomas with S100 protein expression. On the other hand, NF-2 copy number loss was associated with lower SWI signal intensity, a growth pattern described as "en plaque", and the presence of calcification within the tumor. The logistic regression model achieved an accuracy of 0.59 for predicting NF-2 copy number loss and an accuracy of 0.70 for identifying S100 protein expression. Deep learning features demonstrated a strong predictive capability for S100 protein expression (AUC = 0.85 ± 0.06) and had reasonable success in identifying NF-2 copy number loss (AUC = 0.74 ± 0.05). In conclusion, SWI showed promise in identifying NF-2 copy number loss and S100 protein expression by revealing neovascularization and microcalcification characteristics in meningiomas.