Artificial intelligence-driven radiomics study in cancer:the role of feature engineering and modeling

Modern medicine is reliant on various medical imaging technologies for non-invasively observing patients’anatomy.However,the interpretation of medical images can be highly subjective and dependent on the expertise of clinicians.Moreover,some potentially useful quantitative information in medical images,especially that which is not visible to the naked eye,is often ignored during clinical practice.In contrast,radiomics performs high-throughput feature extraction from medical images,which enables quantitative analysis of medical images and prediction of various clinical endpoints.Studies have reported that radiomics exhibits promising performance in diagnosis and predicting treatment responses and prognosis,demonstrating its potential to be a non-invasive auxiliary tool for personalized medicine.However,radiomics remains in a developmental phase as numerous technical challenges have yet to be solved,especially in feature engineering and statistical modeling.In this review,we introduce the current utility of radiomics by summarizing research on its application in the diagnosis,prognosis,and prediction of treatment responses in patients with cancer.We focus on machine learning approaches,for feature extraction and selection during feature engineering and for imbalanced datasets and multi-modality fusion during statistical modeling.Furthermore,we introduce the stability,reproducibility,and interpretability of features,and the generalizability and interpretability of models.Finally,we offer possible solutions to current challenges in radiomics research.

Point Cloud Classification Using Content-Based Transformer via Clustering in Feature Space

Recently, there have been some attempts of Transformer in 3D point cloud classification. In order to reduce computations, most existing methods focus on local spatial attention,but ignore their content and fail to establish relationships between distant but relevant points. To overcome the limitation of local spatial attention, we propose a point content-based Transformer architecture, called PointConT for short. It exploits the locality of points in the feature space(content-based), which clusters the sampled points with similar features into the same class and computes the self-attention within each class, thus enabling an effective trade-off between capturing long-range dependencies and computational complexity. We further introduce an inception feature aggregator for point cloud classification, which uses parallel structures to aggregate high-frequency and low-frequency information in each branch separately. Extensive experiments show that our PointConT model achieves a remarkable performance on point cloud shape classification. Especially, our method exhibits 90.3% Top-1 accuracy on the hardest setting of ScanObjectN N. Source code of this paper is available at https://github.com/yahuiliu99/PointC onT.

Epidemiological and clinical features,treatment status,and economic burden of traumatic spinal cord injury in China:a hospital-based retrospective study

Traumatic spinal cord injury is potentially catastrophic and can lead to permanent disability or even death.China has the largest population of patients with traumatic spinal cord injury.Previous studies of traumatic spinal cord injury in China have mostly been regional in scope;national-level studies have been rare.To the best of our knowledge,no national-level study of treatment status and economic burden has been performed.This retrospective study aimed to examine the epidemiological and clinical features,treatment status,and economic burden of traumatic spinal cord injury in China at the national level.We included 13,465 traumatic spinal cord injury patients who were injured between January 2013 and December 2018 and treated in 30 hospitals in 11 provinces/municipalities representing all geographical divisions of China.Patient epidemiological and clinical features,treatment status,and total and daily costs were recorded.Trends in the percentage of traumatic spinal cord injuries among all hospitalized patients and among patients hospitalized in the orthopedic department and cost of care were assessed by annual percentage change using the Joinpoint Regression Program.The percentage of traumatic spinal cord injuries among all hospitalized patients and among patients hospitalized in the orthopedic department did not significantly change overall(annual percentage change,-0.5%and 2.1%,respectively).A total of 10,053(74.7%)patients underwent surgery.Only 2.8%of patients who underwent surgery did so within 24 hours of injury.A total of 2005(14.9%)patients were treated with high-dose(≥500 mg)methylprednisolone sodium succinate/methylprednisolone(MPSS/MP);615(4.6%)received it within 8 hours.The total cost for acute traumatic spinal cord injury decreased over the study period(-4.7%),while daily cost did not significantly change(1.0%increase).Our findings indicate that public health initiatives should aim at improving hospitals’ability to complete early surgery within 24 hours,which is associated with improved sensorimotor recovery,increasing the awareness rate of clinical guidelines related to high-dose MPSS/MP to reduce the use of the treatment with insufficient evidence.

Endoscopic features and treatments of gastric cystica profunda

BACKGROUND Gastric cystica profunda(GCP)represents a rare condition characterized by cystic dilation of gastric glands within the mucosal and/or submucosal layers.GCP is often linked to,or may progress into,early gastric cancer(EGC).AIM To provide a comprehensive evaluation of the endoscopic features of GCP while assessing the efficacy of endoscopic treatment,thereby offering guidance for diagnosis and treatment.METHODS This retrospective study involved 104 patients with GCP who underwent endoscopic resection.Alongside demographic and clinical data,regular patient followups were conducted to assess local recurrence.RESULTS Among the 104 patients diagnosed with GCP who underwent endoscopic resection,12.5%had a history of previous gastric procedures.The primary site predominantly affected was the cardia(38.5%,n=40).GCP commonly exhibited intraluminal growth(99%),regular presentation(74.0%),and ulcerative mucosa(61.5%).The leading endoscopic feature was the mucosal lesion type(59.6%,n=62).The average maximum diameter was 20.9±15.3 mm,with mucosal involvement in 60.6%(n=63).Procedures lasted 73.9±57.5 min,achieving complete resection in 91.3%(n=95).Recurrence(4.8%)was managed via either surgical intervention(n=1)or through endoscopic resection(n=4).Final pathology confirmed that 59.6%of GCP cases were associated with EGC.Univariate analysis indicated that elderly males were more susceptible to GCP associated with EGC.Conversely,multivariate analysis identified lesion morphology and endoscopic features as significant risk factors.Survival analysis demonstrated no statistically significant difference in recurrence between GCP with and without EGC(P=0.72).CONCLUSION The findings suggested that endoscopic resection might serve as an effective and minimally invasive treatment for GCP with or without EGC.

BDPartNet: Feature Decoupling and Reconstruction Fusion Network for Infrared and Visible Image

While single-modal visible light images or infrared images provide limited information,infrared light captures significant thermal radiation data,whereas visible light excels in presenting detailed texture information.Com-bining images obtained from both modalities allows for leveraging their respective strengths and mitigating individual limitations,resulting in high-quality images with enhanced contrast and rich texture details.Such capabilities hold promising applications in advanced visual tasks including target detection,instance segmentation,military surveillance,pedestrian detection,among others.This paper introduces a novel approach,a dual-branch decomposition fusion network based on AutoEncoder(AE),which decomposes multi-modal features into intensity and texture information for enhanced fusion.Local contrast enhancement module(CEM)and texture detail enhancement module(DEM)are devised to process the decomposed images,followed by image fusion through the decoder.The proposed loss function ensures effective retention of key information from the source images of both modalities.Extensive comparisons and generalization experiments demonstrate the superior performance of our network in preserving pixel intensity distribution and retaining texture details.From the qualitative results,we can see the advantages of fusion details and local contrast.In the quantitative experiments,entropy(EN),mutual information(MI),structural similarity(SSIM)and other results have improved and exceeded the SOTA(State of the Art)model as a whole.

Automatic Extraction Method of 3D Feature Guidelines for Complex Cultural Relic Surfaces Based on Point Cloud

Cultural relics line graphic serves as a crucial form of traditional artifact information documentation,which is a simple and intuitive product with low cost of displaying compared with 3D models.Dimensionality reduction is undoubtedly necessary for line drawings.However,most existing methods for artifact drawing rely on the principles of orthographic projection that always cannot avoid angle occlusion and data overlapping while the surface of cultural relics is complex.Therefore,conformal mapping was introduced as a dimensionality reduction way to compensate for the limitation of orthographic projection.Based on the given criteria for assessing surface complexity,this paper proposed a three-dimensional feature guideline extraction method for complex cultural relic surfaces.A 2D and 3D combined factor that measured the importance of points on describing surface features,vertex weight,was designed.Then the selection threshold for feature guideline extraction was determined based on the differences between vertex weight and shape index distributions.The feasibility and stability were verified through experiments conducted on real cultural relic surface data.Results demonstrated the ability of the method to address the challenges associated with the automatic generation of line drawings for complex surfaces.The extraction method and the obtained results will be useful for line graphic drawing,displaying and propaganda of cultural relics.

A Model for Detecting Fake News by Integrating Domain-Specific Emotional and Semantic Features

With the rapid spread of Internet information and the spread of fake news,the detection of fake news becomes more and more important.Traditional detection methods often rely on a single emotional or semantic feature to identify fake news,but these methods have limitations when dealing with news in specific domains.In order to solve the problem of weak feature correlation between data from different domains,a model for detecting fake news by integrating domain-specific emotional and semantic features is proposed.This method makes full use of the attention mechanism,grasps the correlation between different features,and effectively improves the effect of feature fusion.The algorithm first extracts the semantic features of news text through the Bi-LSTM(Bidirectional Long Short-Term Memory)layer to capture the contextual relevance of news text.Senta-BiLSTM is then used to extract emotional features and predict the probability of positive and negative emotions in the text.It then uses domain features as an enhancement feature and attention mechanism to fully capture more fine-grained emotional features associated with that domain.Finally,the fusion features are taken as the input of the fake news detection classifier,combined with the multi-task representation of information,and the MLP and Softmax functions are used for classification.The experimental results show that on the Chinese dataset Weibo21,the F1 value of this model is 0.958,4.9% higher than that of the sub-optimal model;on the English dataset FakeNewsNet,the F1 value of the detection result of this model is 0.845,1.8% higher than that of the sub-optimal model,which is advanced and feasible.

MSD-Net: Pneumonia Classification Model Based on Multi-Scale Directional Feature Enhancement

Computer-aided diagnosis of pneumonia based on deep learning is a research hotspot.However,there are some problems that the features of different sizes and different directions are not sufficient when extracting the features in lung X-ray images.A pneumonia classification model based on multi-scale directional feature enhancement MSD-Net is proposed in this paper.The main innovations are as follows:Firstly,the Multi-scale Residual Feature Extraction Module(MRFEM)is designed to effectively extract multi-scale features.The MRFEM uses dilated convolutions with different expansion rates to increase the receptive field and extract multi-scale features effectively.Secondly,the Multi-scale Directional Feature Perception Module(MDFPM)is designed,which uses a three-branch structure of different sizes convolution to transmit direction feature layer by layer,and focuses on the target region to enhance the feature information.Thirdly,the Axial Compression Former Module(ACFM)is designed to perform global calculations to enhance the perception ability of global features in different directions.To verify the effectiveness of the MSD-Net,comparative experiments and ablation experiments are carried out.In the COVID-19 RADIOGRAPHY DATABASE,the Accuracy,Recall,Precision,F1 Score,and Specificity of MSD-Net are 97.76%,95.57%,95.52%,95.52%,and 98.51%,respectively.In the chest X-ray dataset,the Accuracy,Recall,Precision,F1 Score and Specificity of MSD-Net are 97.78%,95.22%,96.49%,95.58%,and 98.11%,respectively.This model improves the accuracy of lung image recognition effectively and provides an important clinical reference to pneumonia Computer-Aided Diagnosis.

Japanese Sign Language Recognition by Combining Joint Skeleton-Based Handcrafted and Pixel-Based Deep Learning Features with Machine Learning Classification

Sign language recognition is vital for enhancing communication accessibility among the Deaf and hard-of-hearing communities.In Japan,approximately 360,000 individualswith hearing and speech disabilities rely on Japanese Sign Language(JSL)for communication.However,existing JSL recognition systems have faced significant performance limitations due to inherent complexities.In response to these challenges,we present a novel JSL recognition system that employs a strategic fusion approach,combining joint skeleton-based handcrafted features and pixel-based deep learning features.Our system incorporates two distinct streams:the first stream extracts crucial handcrafted features,emphasizing the capture of hand and body movements within JSL gestures.Simultaneously,a deep learning-based transfer learning stream captures hierarchical representations of JSL gestures in the second stream.Then,we concatenated the critical information of the first stream and the hierarchy of the second stream features to produce the multiple levels of the fusion features,aiming to create a comprehensive representation of the JSL gestures.After reducing the dimensionality of the feature,a feature selection approach and a kernel-based support vector machine(SVM)were used for the classification.To assess the effectiveness of our approach,we conducted extensive experiments on our Lab JSL dataset and a publicly available Arabic sign language(ArSL)dataset.Our results unequivocally demonstrate that our fusion approach significantly enhances JSL recognition accuracy and robustness compared to individual feature sets or traditional recognition methods.

Constitution identification model in traditional Chinese medicine based on multiple features

Objective To construct a precise model for identifying traditional Chinese medicine(TCM)constitutions;thereby offering optimized guidance for clinical diagnosis and treatment plan-ning;and ultimately enhancing medical efficiency and treatment outcomes.Methods First;TCM full-body inspection data acquisition equipment was employed to col-lect full-body standing images of healthy people;from which the constitutions were labelled and defined in accordance with the Constitution in Chinese Medicine Questionnaire(CCMQ);and a dataset encompassing labelled constitutions was constructed.Second;heat-suppres-sion valve(HSV)color space and improved local binary patterns(LBP)algorithm were lever-aged for the extraction of features such as facial complexion and body shape.In addition;a dual-branch deep network was employed to collect deep features from the full-body standing images.Last;the random forest(RF)algorithm was utilized to learn the extracted multifea-tures;which were subsequently employed to establish a TCM constitution identification mod-el.Accuracy;precision;and F1 score were the three measures selected to assess the perfor-mance of the model.Results It was found that the accuracy;precision;and F1 score of the proposed model based on multifeatures for identifying TCM constitutions were 0.842;0.868;and 0.790;respectively.In comparison with the identification models that encompass a single feature;either a single facial complexion feature;a body shape feature;or deep features;the accuracy of the model that incorporating all the aforementioned features was elevated by 0.105;0.105;and 0.079;the precision increased by 0.164;0.164;and 0.211;and the F1 score rose by 0.071;0.071;and 0.084;respectively.Conclusion The research findings affirmed the viability of the proposed model;which incor-porated multifeatures;including the facial complexion feature;the body shape feature;and the deep feature.In addition;by employing the proposed model;the objectification and intel-ligence of identifying constitutions in TCM practices could be optimized.

Cross-Dimension Attentive Feature Fusion Network for Unsupervised Time-Series Anomaly Detection

Time series anomaly detection is crucial in various industrial applications to identify unusual behaviors within the time series data.Due to the challenges associated with annotating anomaly events,time series reconstruction has become a prevalent approach for unsupervised anomaly detection.However,effectively learning representations and achieving accurate detection results remain challenging due to the intricate temporal patterns and dependencies in real-world time series.In this paper,we propose a cross-dimension attentive feature fusion network for time series anomaly detection,referred to as CAFFN.Specifically,a series and feature mixing block is introduced to learn representations in 1D space.Additionally,a fast Fourier transform is employed to convert the time series into 2D space,providing the capability for 2D feature extraction.Finally,a cross-dimension attentive feature fusion mechanism is designed that adaptively integrates features across different dimensions for anomaly detection.Experimental results on real-world time series datasets demonstrate that CAFFN performs better than other competing methods in time series anomaly detection.

FusionNN:A Semantic Feature Fusion Model Based on Multimodal for Web Anomaly Detection

With the rapid development of the mobile communication and the Internet,the previous web anomaly detectionand identificationmodels were built relying on security experts’empirical knowledge and attack features.Althoughthis approach can achieve higher detection performance,it requires huge human labor and resources to maintainthe feature library.In contrast,semantic feature engineering can dynamically discover new semantic featuresand optimize feature selection by automatically analyzing the semantic information contained in the data itself,thus reducing dependence on prior knowledge.However,current semantic features still have the problem ofsemantic expression singularity,as they are extracted from a single semantic mode such as word segmentation,character segmentation,or arbitrary semantic feature extraction.This paper extracts features of web requestsfrom dual semantic granularity,and proposes a semantic feature fusion method to solve the above problems.Themethod first preprocesses web requests,and extracts word-level and character-level semantic features of URLs viaconvolutional neural network(CNN),respectively.By constructing three loss functions to reduce losses betweenfeatures,labels and categories.Experiments on the HTTP CSIC 2010,Malicious URLs and HttpParams datasetsverify the proposedmethod.Results show that compared withmachine learning,deep learningmethods and BERTmodel,the proposed method has better detection performance.And it achieved the best detection rate of 99.16%in the dataset HttpParams.

Three-dimensional Modeling of Ore-forming Elements and Mineralization Prognosis for the Yechangping Mo Deposit,Henan Province,China

Three-dimensional geochemical modeling of ore-forming elements is crucial for predicting deep mineralization.This approach provides key information for the quantitative prediction of deep mineral localization,three-dimensional fine interpolation,analysis of spatial distribution patterns,and extraction of quantitative mineral-seeking markers.The Yechangping molybdenum(Mo)deposit is a significant and extensive porphyry-skarn deposit in the East Qinling-Dabie Mo polymetallic metallogenic belt at the southern margin of the North China Block.Abundant borehole data on oreforming elements underpin deep geochemical predictions.The methodology includes the following steps:(1)Threedimensional geological modeling of the deposit was established.(2)Correlation,cluster,and factor analyses post delineation of mineralization stages and determination of mineral generation sequence to identify(Cu,Pb,Zn,Ag)and(Mo,W,mfe)assemblages.(3)A three-dimensional geochemical block model was constructed for Mo,W,mfe,Cu,Zn,Pb,and Ag using the ordinary kriging method,and the variational function was developed.(4)Spatial distribution and enrichment characteristics analysis of ore-forming elements are performed to extract geological information,employing the variogram and w(Cu+Pb+Zn+Ag)/w(Mo+W)as predictive indicators.(5)Identifying the western,northwestern,and southwestern areas of the mine with limited mineralization potential,contrasted by the northeastern and southeastern areas favorable for mineral exploration.

Video-Based Deception Detection with Non-Contact Heart Rate Monitoring and Multi-Modal Feature Selection

Deception detection plays a crucial role in criminal investigation.Videos contain a wealth of information regarding apparent and physiological changes in individuals,and thus can serve as an effective means of deception detection.In this paper,we investigate video-based deception detection considering both apparent visual features such as eye gaze,head pose and facial action unit(AU),and non-contact heart rate detected by remote photoplethysmography(rPPG)technique.Multiple wrapper-based feature selection methods combined with the K-nearest neighbor(KNN)and support vector machine(SVM)classifiers are employed to screen the most effective features for deception detection.We evaluate the performance of the proposed method on both a self-collected physiological-assisted visual deception detection(PV3D)dataset and a public bag-oflies(BOL)dataset.Experimental results demonstrate that the SVM classifier with symbiotic organisms search(SOS)feature selection yields the best overall performance,with an area under the curve(AUC)of 83.27%and accuracy(ACC)of 83.33%for PV3D,and an AUC of 71.18%and ACC of 70.33%for BOL.This demonstrates the stability and effectiveness of the proposed method in video-based deception detection tasks.

GP‐FMLNet:A feature matrix learning network enhanced by glyph and phonetic information for Chinese sentiment analysis

Sentiment analysis is a fine‐grained analysis task that aims to identify the sentiment polarity of a specified sentence.Existing methods in Chinese sentiment analysis tasks only consider sentiment features from a single pole and scale and thus cannot fully exploit and utilise sentiment feature information,making their performance less than ideal.To resolve the problem,the authors propose a new method,GP‐FMLNet,that integrates both glyph and phonetic information and design a novel feature matrix learning process for phonetic features with which to model words that have the same pinyin information but different glyph information.Our method solves the problem of misspelling words influencing sentiment polarity prediction results.Specifically,the authors iteratively mine character,glyph,and pinyin features from the input comments sentences.Then,the authors use soft attention and matrix compound modules to model the phonetic features,which empowers their model to keep on zeroing in on the dynamic‐setting words in various positions and to dispense with the impacts of the deceptive‐setting ones.Ex-periments on six public datasets prove that the proposed model fully utilises the glyph and phonetic information and improves on the performance of existing Chinese senti-ment analysis algorithms.

A Heuristic Radiomics Feature Selection Method Based on Frequency Iteration and Multi-Supervised Training Mode

Radiomics is a non-invasive method for extracting quantitative and higher-dimensional features from medical images for diagnosis.It has received great attention due to its huge application prospects in recent years.We can know that the number of features selected by the existing radiomics feature selectionmethods is basically about ten.In this paper,a heuristic feature selection method based on frequency iteration and multiple supervised training mode is proposed.Based on the combination between features,it decomposes all features layer by layer to select the optimal features for each layer,then fuses the optimal features to form a local optimal group layer by layer and iterates to the global optimal combination finally.Compared with the currentmethod with the best prediction performance in the three data sets,thismethod proposed in this paper can reduce the number of features fromabout ten to about three without losing classification accuracy and even significantly improving classification accuracy.The proposed method has better interpretability and generalization ability,which gives it great potential in the feature selection of radiomics.

Feature extraction and learning approaches for cancellable biometrics:A survey

Biometric recognition is a widely used technology for user authentication.In the application of this technology,biometric security and recognition accuracy are two important issues that should be considered.In terms of biometric security,cancellable biometrics is an effective technique for protecting biometric data.Regarding recognition accuracy,feature representation plays a significant role in the performance and reliability of cancellable biometric systems.How to design good feature representations for cancellable biometrics is a challenging topic that has attracted a great deal of attention from the computer vision community,especially from researchers of cancellable biometrics.Feature extraction and learning in cancellable biometrics is to find suitable feature representations with a view to achieving satisfactory recognition performance,while the privacy of biometric data is protected.This survey informs the progress,trend and challenges of feature extraction and learning for cancellable biometrics,thus shedding light on the latest developments and future research of this area.

Curve Classification Based onMean-Variance Feature Weighting and Its Application

The classification of functional data has drawn much attention in recent years.The main challenge is representing infinite-dimensional functional data by finite-dimensional features while utilizing those features to achieve better classification accuracy.In this paper,we propose a mean-variance-based(MV)feature weighting method for classifying functional data or functional curves.In the feature extraction stage,each sample curve is approximated by B-splines to transfer features to the coefficients of the spline basis.After that,a feature weighting approach based on statistical principles is introduced by comprehensively considering the between-class differences and within-class variations of the coefficients.We also introduce a scaling parameter to adjust the gap between the weights of features.The new feature weighting approach can adaptively enhance noteworthy local features while mitigating the impact of confusing features.The algorithms for feature weighted K-nearest neighbor and support vector machine classifiers are both provided.Moreover,the new approach can be well integrated into existing functional data classifiers,such as the generalized functional linear model and functional linear discriminant analysis,resulting in a more accurate classification.The performance of the mean-variance-based classifiers is evaluated by simulation studies and real data.The results show that the newfeatureweighting approach significantly improves the classification accuracy for complex functional data.

Feature extraction for machine learning-based intrusion detection in IoT networks

A large number of network security breaches in IoT networks have demonstrated the unreliability of current Network Intrusion Detection Systems(NIDSs).Consequently,network interruptions and loss of sensitive data have occurred,which led to an active research area for improving NIDS technologies.In an analysis of related works,it was observed that most researchers aim to obtain better classification results by using a set of untried combinations of Feature Reduction(FR)and Machine Learning(ML)techniques on NIDS datasets.However,these datasets are different in feature sets,attack types,and network design.Therefore,this paper aims to discover whether these techniques can be generalised across various datasets.Six ML models are utilised:a Deep Feed Forward(DFF),Convolutional Neural Network(CNN),Recurrent Neural Network(RNN),Decision Tree(DT),Logistic Regression(LR),and Naive Bayes(NB).The accuracy of three Feature Extraction(FE)algorithms is detected;Principal Component Analysis(PCA),Auto-encoder(AE),and Linear Discriminant Analysis(LDA),are evaluated using three benchmark datasets:UNSW-NB15,ToN-IoT and CSE-CIC-IDS2018.Although PCA and AE algorithms have been widely used,the determination of their optimal number of extracted dimensions has been overlooked.The results indicate that no clear FE method or ML model can achieve the best scores for all datasets.The optimal number of extracted dimensions has been identified for each dataset,and LDA degrades the performance of the ML models on two datasets.The variance is used to analyse the extracted dimensions of LDA and PCA.Finally,this paper concludes that the choice of datasets significantly alters the performance of the applied techniques.We believe that a universal(benchmark)feature set is needed to facilitate further advancement and progress of research in this field.

Research on Multi-Scale Feature Fusion Network Algorithm Based on Brain Tumor Medical Image Classification

Gliomas have the highest mortality rate of all brain tumors.Correctly classifying the glioma risk period can help doctors make reasonable treatment plans and improve patients’survival rates.This paper proposes a hierarchical multi-scale attention feature fusion medical image classification network(HMAC-Net),which effectively combines global features and local features.The network framework consists of three parallel layers:The global feature extraction layer,the local feature extraction layer,and the multi-scale feature fusion layer.A linear sparse attention mechanism is designed in the global feature extraction layer to reduce information redundancy.In the local feature extraction layer,a bilateral local attention mechanism is introduced to improve the extraction of relevant information between adjacent slices.In the multi-scale feature fusion layer,a channel fusion block combining convolutional attention mechanism and residual inverse multi-layer perceptron is proposed to prevent gradient disappearance and network degradation and improve feature representation capability.The double-branch iterative multi-scale classification block is used to improve the classification performance.On the brain glioma risk grading dataset,the results of the ablation experiment and comparison experiment show that the proposed HMAC-Net has the best performance in both qualitative analysis of heat maps and quantitative analysis of evaluation indicators.On the dataset of skin cancer classification,the generalization experiment results show that the proposed HMAC-Net has a good generalization effect.