Smart Health最新文献_第7页

Human knowledge-based artificial intelligence methods for skin cancer management: Accuracy and interpretability study 基于人类知识的皮肤癌管理人工智能方法：准确性和可解释性研究

Q2 Health Professions

Smart Health

Pub Date : 2025-01-23 DOI: 10.1016/j.smhl.2025.100540

Eman Rezk , Mohamed Eltorki , Wael El-Dakhakhni

Skin cancer management, including monitoring and excision, involves sophisticated decisions reliant on several interdependent factors. This complexity leads to a scarcity of data useful for skin cancer management. Deep learning achieved massive success in computer vision due to its ability to extract representative features from images. However, deep learning methods require large amounts of data to develop accurate models, whereas machine learning methods perform well with small datasets. In this work, we aim to compare the accuracy and interpretability of skin cancer management prediction 1) using deep learning and machine learning methods and 2) utilizing various inputs including clinical images, dermoscopic images, and lesion clinical tabular features created by experts to represent lesion characteristics. We implemented two approaches, a deep learning pipeline for feature extraction and classification trained on different input modalities including images and lesion clinical features. The second approach uses lesion clinical features to train machine learning classifiers. The results show that the machine learning approach trained on clinical features achieves higher accuracy (0.80) and higher area under the curve (0.92) compared to the deep learning pipeline trained on skin images and lesion clinical features which achieves an accuracy of 0.66 and area under the curve of 0.74. Additionally, the machine learning approach provides more informative and understandable interpretations of the results. This work emphasizes the significance of utilizing human knowledge in developing precise and transparent predictive models. In addition, our findings highlight the potential of machine learning methods in predicting lesion management in situation where the data size is insufficient to leverage deep learning capabilities.

皮肤癌的管理，包括监测和切除，涉及依赖于几个相互依存因素的复杂决策。这种复杂性导致对皮肤癌管理有用的数据缺乏。由于能够从图像中提取代表性特征，深度学习在计算机视觉领域取得了巨大的成功。然而，深度学习方法需要大量的数据来开发准确的模型，而机器学习方法在小数据集上表现良好。在这项工作中，我们的目标是比较皮肤癌管理预测的准确性和可解释性：1)使用深度学习和机器学习方法；2)利用各种输入，包括临床图像、皮肤镜图像和专家创建的病变临床表格特征来代表病变特征。我们实现了两种方法，一种是深度学习管道，用于特征提取和分类，训练于不同的输入模式，包括图像和病变临床特征。第二种方法使用病变临床特征来训练机器学习分类器。结果表明，与基于皮肤图像和病变临床特征训练的深度学习管道相比，基于临床特征训练的机器学习方法获得了更高的准确率（0.80）和曲线下面积（0.92），后者的准确率为0.66，曲线下面积为0.74。此外，机器学习方法为结果提供了更多信息和可理解的解释。这项工作强调了利用人类知识开发精确和透明的预测模型的重要性。此外，我们的研究结果强调了机器学习方法在数据量不足以利用深度学习能力的情况下预测病变管理的潜力。

{"title":"Human knowledge-based artificial intelligence methods for skin cancer management: Accuracy and interpretability study","authors":"Eman Rezk , Mohamed Eltorki , Wael El-Dakhakhni","doi":"10.1016/j.smhl.2025.100540","DOIUrl":"10.1016/j.smhl.2025.100540","url":null,"abstract":"<div><div>Skin cancer management, including monitoring and excision, involves sophisticated decisions reliant on several interdependent factors. This complexity leads to a scarcity of data useful for skin cancer management. Deep learning achieved massive success in computer vision due to its ability to extract representative features from images. However, deep learning methods require large amounts of data to develop accurate models, whereas machine learning methods perform well with small datasets. In this work, we aim to compare the accuracy and interpretability of skin cancer management prediction 1) using deep learning and machine learning methods and 2) utilizing various inputs including clinical images, dermoscopic images, and lesion clinical tabular features created by experts to represent lesion characteristics. We implemented two approaches, a deep learning pipeline for feature extraction and classification trained on different input modalities including images and lesion clinical features. The second approach uses lesion clinical features to train machine learning classifiers. The results show that the machine learning approach trained on clinical features achieves higher accuracy (0.80) and higher area under the curve (0.92) compared to the deep learning pipeline trained on skin images and lesion clinical features which achieves an accuracy of 0.66 and area under the curve of 0.74. Additionally, the machine learning approach provides more informative and understandable interpretations of the results. This work emphasizes the significance of utilizing human knowledge in developing precise and transparent predictive models. In addition, our findings highlight the potential of machine learning methods in predicting lesion management in situation where the data size is insufficient to leverage deep learning capabilities.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"36 ","pages":"Article 100540"},"PeriodicalIF":0.0,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143100542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

SAFE: Sound Analysis for Fall Event detection using machine learning SAFE：使用机器学习进行坠落事件检测的可靠分析

Q2 Health Professions

Smart Health

Pub Date : 2025-01-06 DOI: 10.1016/j.smhl.2024.100539

Antony Garcia , Xinming Huang

This study evaluates the application of machine learning (ML) and deep learning (DL) algorithms for fall detection using sound signals. The work is supported by the Sound Analysis for Fall Events (SAFE) dataset, comprising 950 audio samples, including 475 fall events recorded with a grappling dummy to simulate realistic scenarios. Decision tree-based ML algorithms achieved a classification accuracy of 93% at lower sampling rates, indicating that critical features are preserved despite reduced resolution. DL models, using spectrogram-based feature extraction, reached accuracies up to 99%, surpassing traditional ML methods in performance. Linear models also achieved high accuracy (up to 97%) in various spectrogram techniques, emphasizing the separability of audio features. These results establish the viability of sound-based fall detection systems as efficient and accurate solutions.

本研究评估了机器学习（ML）和深度学习（DL）算法在使用声音信号进行跌倒检测中的应用。这项工作得到了跌落事件声音分析（SAFE）数据集的支持，该数据集由950个音频样本组成，其中包括475个跌落事件，这些事件是用抓握假人录制的，以模拟现实场景。基于决策树的机器学习算法在较低的采样率下实现了93%的分类准确率，这表明尽管分辨率降低了，但仍然保留了关键特征。DL模型使用基于谱图的特征提取，准确率高达99%，在性能上超过了传统的ML方法。线性模型在各种谱图技术中也实现了高精度（高达97%），强调了音频特征的可分离性。这些结果表明，基于声音的坠落检测系统是一种高效、准确的解决方案。

引用次数: 0

Latent Space Representation of Adversarial AutoEncoder for Human Activity Recognition: Application to a low-cost commercial force plate and inertial measurement units 用于人体活动识别的对抗性自编码器的潜在空间表示：在低成本商用测力板和惯性测量单元上的应用

Q2 Health Professions

Smart Health

Pub Date : 2025-01-04 DOI: 10.1016/j.smhl.2024.100537

Kenta Kamikokuryo , Gentiane Venture , Vincent Hernandez

Human Activity Recognition (HAR) is a key component of a home rehabilitation system that provides real-time monitoring and personalized feedback. This research explores the application of Adversarial AutoEncoder (AAE) models for data dimensionality reduction in the context of HAR. Visualizing data in a lower-dimensional space is important to understand changes in motor control due to medical conditions or aging, to aid personalized interventions, and to ensure continuous benefits in remote rehabilitation settings. This makes patient assessment effective, easier, and faster.

In this study, the classification performance of the latent space created by the AAE is evaluated using the Wii Balance Board (WiiBB) and/or three Inertial Measurement Units (IMUs) placed on the forearms and hip. Various sensor configurations are considered, including only WiiBB, only IMUs, combinations of WiiBB with the IMU at the hip, and combinations of WiiBB with the 3 IMUs.

The accuracy of the latent space representation is compared with two common supervised classification models, which are the Convolutional Neural Network (CNN) and the neural network called CNNLSTM, which is composed of convolution layers followed by recurrent layers. The approach was demonstrated for two different sets of exercises consisting of upper and lower body exercises collected with 19 participants.

The results show that the latent space representation of the AAE achieves a strong classification accuracy performance while also serving as a visualization tool. This study is an initial demonstration of the potential of integrating WiiBB and IMU sensors for comprehensive activity recognition for upper and lower body movement analysis.

人类活动识别（HAR）是提供实时监测和个性化反馈的家庭康复系统的关键组成部分。本研究探讨了对抗自动编码器（AAE）模型在HAR背景下的数据降维应用。在低维空间中可视化数据对于了解由于医疗条件或衰老导致的运动控制变化，帮助个性化干预以及确保远程康复环境中的持续效益非常重要。这使得对病人的评估更有效、更容易和更快。在本研究中，使用Wii平衡板（WiiBB）和/或放置在前臂和臀部的三个惯性测量单元（imu）来评估由AAE产生的潜在空间的分类性能。考虑了各种传感器配置，包括仅WiiBB，仅IMU， WiiBB与臀部IMU的组合，以及WiiBB与3个IMU的组合。将潜在空间表示的准确性与卷积神经网络（CNN）和由卷积层和递归层组成的CNNLSTM神经网络两种常见的监督分类模型进行了比较。该方法通过收集19名参与者的上半身和下半身练习两组不同的练习进行了演示。结果表明，隐空间表示在获得较强分类精度的同时，还可以作为一种可视化工具。这项研究初步证明了将WiiBB和IMU传感器集成到上半身和下半身运动分析的综合活动识别中的潜力。

{"title":"Latent Space Representation of Adversarial AutoEncoder for Human Activity Recognition: Application to a low-cost commercial force plate and inertial measurement units","authors":"Kenta Kamikokuryo , Gentiane Venture , Vincent Hernandez","doi":"10.1016/j.smhl.2024.100537","DOIUrl":"10.1016/j.smhl.2024.100537","url":null,"abstract":"<div><div>Human Activity Recognition (HAR) is a key component of a home rehabilitation system that provides real-time monitoring and personalized feedback. This research explores the application of Adversarial AutoEncoder (AAE) models for data dimensionality reduction in the context of HAR. Visualizing data in a lower-dimensional space is important to understand changes in motor control due to medical conditions or aging, to aid personalized interventions, and to ensure continuous benefits in remote rehabilitation settings. This makes patient assessment effective, easier, and faster.</div><div>In this study, the classification performance of the latent space created by the AAE is evaluated using the Wii Balance Board (WiiBB) and/or three Inertial Measurement Units (IMUs) placed on the forearms and hip. Various sensor configurations are considered, including only WiiBB, only IMUs, combinations of WiiBB with the IMU at the hip, and combinations of WiiBB with the 3 IMUs.</div><div>The accuracy of the latent space representation is compared with two common supervised classification models, which are the Convolutional Neural Network (CNN) and the neural network called CNNLSTM, which is composed of convolution layers followed by recurrent layers. The approach was demonstrated for two different sets of exercises consisting of upper and lower body exercises collected with 19 participants.</div><div>The results show that the latent space representation of the AAE achieves a strong classification accuracy performance while also serving as a visualization tool. This study is an initial demonstration of the potential of integrating WiiBB and IMU sensors for comprehensive activity recognition for upper and lower body movement analysis.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100537"},"PeriodicalIF":0.0,"publicationDate":"2025-01-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129614","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Novel EEG feature selection based on hellinger distance for epileptic seizure detection 基于hellinger距离的脑电图特征选择在癫痫发作检测中的应用

Q2 Health Professions

Smart Health

Pub Date : 2025-01-01 DOI: 10.1016/j.smhl.2024.100536

Muhammed Sadiq , Mustafa Noaman Kadhim , Dhiah Al-Shammary , Mariofanna Milanova

This study introduces a novel feature selection method based on Hellinger distance and particle swarm optimization (PSO) for reducing the dimensionality of features in electroencephalogram (EEG) signals and improving epileptic seizure detection accuracy. In the first phase, the Hellinger distance is used as a filter to remove redundant and irrelevant features by calculating the similarity between blocks within the feature, thus reducing the search space for the subsequent second phase. In the second phase, PSO searches the reduced feature space to select the best subset. Recognizing that both classification accuracy and dimensionality play crucial roles in the performance of feature subsets, PSO searches various sets of features (ranging from 410 to 2867 in EEG signals) derived from the first stage using Hellinger distance, rather than searching through the full set of 4047 features, to select the optimal subset. The proposed Hellinger-PSO approach demonstrates significant improvements in classification accuracy across multiple models. Specifically, Logistic Regression (LR) improved from 91% to 95% (4% improvement), Decision Tree (DT) from 95% to 97% (2% improvement), Naive Bayes (NB) from 94% to 99% (5% improvement), and Random Forest (RF) from 96% to 98% (2% improvement) on the Bonn dataset. Additionally, the method reduces dimensionality while maintaining high classification performance. The results validate the efficacy of the Hellinger-PSO technique, which enhances both the accuracy and efficiency of epileptic seizure detection. This approach has the potential to improve diagnostic accuracy in medical settings, aiding in better patient care and more effective clinical decision-making.

提出了一种基于海灵格距离和粒子群优化（PSO）的特征选择方法，以降低脑电图（EEG）信号中特征的维数，提高癫痫发作检测的准确率。在第一阶段，使用Hellinger距离作为过滤器，通过计算特征内块之间的相似度来去除冗余和不相关的特征，从而减少后续第二阶段的搜索空间。第二阶段，粒子群算法搜索约简后的特征空间，选择最优子集。认识到分类精度和维数对特征子集的性能起着至关重要的作用，PSO使用海灵格距离搜索从第一阶段得到的各种特征集（EEG信号中的410到2867），而不是搜索4047个特征的全部集合，以选择最优子集。提出的Hellinger-PSO方法在跨多个模型的分类精度方面有显著提高。具体来说，波恩数据集上的逻辑回归（LR）从91%提高到95%（提高4%），决策树（DT）从95%提高到97%（提高2%），朴素贝叶斯（NB）从94%提高到99%（提高5%），随机森林（RF）从96%提高到98%（提高2%）。此外，该方法在保持高分类性能的同时降低了维数。实验结果验证了Hellinger-PSO技术的有效性，提高了癫痫发作检测的准确性和效率。这种方法有可能提高医疗环境中的诊断准确性，帮助更好的患者护理和更有效的临床决策。

{"title":"Novel EEG feature selection based on hellinger distance for epileptic seizure detection","authors":"Muhammed Sadiq , Mustafa Noaman Kadhim , Dhiah Al-Shammary , Mariofanna Milanova","doi":"10.1016/j.smhl.2024.100536","DOIUrl":"10.1016/j.smhl.2024.100536","url":null,"abstract":"<div><div>This study introduces a novel feature selection method based on Hellinger distance and particle swarm optimization (PSO) for reducing the dimensionality of features in electroencephalogram (EEG) signals and improving epileptic seizure detection accuracy. In the first phase, the Hellinger distance is used as a filter to remove redundant and irrelevant features by calculating the similarity between blocks within the feature, thus reducing the search space for the subsequent second phase. In the second phase, PSO searches the reduced feature space to select the best subset. Recognizing that both classification accuracy and dimensionality play crucial roles in the performance of feature subsets, PSO searches various sets of features (ranging from 410 to 2867 in EEG signals) derived from the first stage using Hellinger distance, rather than searching through the full set of 4047 features, to select the optimal subset. The proposed Hellinger-PSO approach demonstrates significant improvements in classification accuracy across multiple models. Specifically, Logistic Regression (LR) improved from 91% to 95% (4% improvement), Decision Tree (DT) from 95% to 97% (2% improvement), Naive Bayes (NB) from 94% to 99% (5% improvement), and Random Forest (RF) from 96% to 98% (2% improvement) on the Bonn dataset. Additionally, the method reduces dimensionality while maintaining high classification performance. The results validate the efficacy of the Hellinger-PSO technique, which enhances both the accuracy and efficiency of epileptic seizure detection. This approach has the potential to improve diagnostic accuracy in medical settings, aiding in better patient care and more effective clinical decision-making.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100536"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129218","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Explainable screening of oral cancer via deep learning and case-based reasoning 通过深度学习和基于案例的推理对口腔癌进行可解释的筛查

Q2 Health Professions

Smart Health

Pub Date : 2025-01-01 DOI: 10.1016/j.smhl.2024.100538

Mario G.C.A. Cimino , Giuseppina Campisi , Federico A. Galatolo , Paolo Neri , Pietro Tozzo , Marco Parola , Gaetano La Mantia , Olga Di Fede

Oral Squamous Cell Carcinoma is characterized by significant mortality and morbidity. Dental professionals can play an important role in its early detection, thanks to the availability of embedded smart cameras for oral photos and remote screening supported by Deep Learning (DL). Despite the promising results of DL for automated detection and classification of oral lesions, its effectiveness is based on a clearly defined protocol, on the explainability of results, and on periodic cases collection. This paper proposes a novel method, combining DL and Case-Based Reasoning (CBR), to allow the post-hoc explanation of the system answer. The method uses explainability tools organized in a protocol defined in the Business Process Model and Notation (BPMN) to allow its experimental validation. A redesign of the Faster-R-CNN Feature Pyramid Networks (FPN) + DL architecture is also proposed for lesions detection and classification, fine-tuned on 160 cases belonging to three classes of oral ulcers. The DL system achieves state-of-the-art performance, i.e., 83% detection and 92% classification rate (98% for neoplastic vs. non-neoplastic binary classification). A preliminary experimentation of the protocol involved both resident and specialized doctors over selected difficult cases. The system and cases have been publicly released to foster collaboration between research centers.

口腔鳞状细胞癌具有显著的死亡率和发病率。由于可用于口腔照片的嵌入式智能摄像头和深度学习（DL）支持的远程筛查，牙科专业人员可以在早期检测中发挥重要作用。尽管DL在口腔病变的自动检测和分类方面有很好的结果，但其有效性是基于明确定义的方案、结果的可解释性和定期病例收集。本文提出了一种新的方法，结合深度学习和基于案例的推理（CBR），允许对系统答案进行事后解释。该方法使用在业务流程模型和符号（BPMN）中定义的协议中组织的可解释性工具来进行实验验证。本文还提出了对Faster-R-CNN特征金字塔网络(FPN) + DL架构的重新设计，用于病灶检测和分类，并对属于三类口腔溃疡的160例病例进行了微调。DL系统达到了最先进的性能，即83%的检测率和92%的分类率（肿瘤与非肿瘤二元分类为98%）。该方案的初步实验涉及住院医生和专科医生对选定的疑难病例。该系统和案例已经公开发布，以促进研究中心之间的合作。

{"title":"Explainable screening of oral cancer via deep learning and case-based reasoning","authors":"Mario G.C.A. Cimino , Giuseppina Campisi , Federico A. Galatolo , Paolo Neri , Pietro Tozzo , Marco Parola , Gaetano La Mantia , Olga Di Fede","doi":"10.1016/j.smhl.2024.100538","DOIUrl":"10.1016/j.smhl.2024.100538","url":null,"abstract":"<div><div>Oral Squamous Cell Carcinoma is characterized by significant mortality and morbidity. Dental professionals can play an important role in its early detection, thanks to the availability of embedded smart cameras for oral photos and remote screening supported by Deep Learning (DL). Despite the promising results of DL for automated detection and classification of oral lesions, its effectiveness is based on a clearly defined protocol, on the explainability of results, and on periodic cases collection. This paper proposes a novel method, combining DL and Case-Based Reasoning (CBR), to allow the post-hoc explanation of the system answer. The method uses explainability tools organized in a protocol defined in the Business Process Model and Notation (BPMN) to allow its experimental validation. A redesign of the Faster-R-CNN Feature Pyramid Networks (FPN) + DL architecture is also proposed for lesions detection and classification, fine-tuned on 160 cases belonging to three classes of oral ulcers. The DL system achieves state-of-the-art performance, i.e., 83% detection and 92% classification rate (98% for neoplastic vs. non-neoplastic binary classification). A preliminary experimentation of the protocol involved both resident and specialized doctors over selected difficult cases. The system and cases have been publicly released to foster collaboration between research centers.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100538"},"PeriodicalIF":0.0,"publicationDate":"2025-01-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129615","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A novel convolutional interpretability model for pixel-level interpretation of medical image classification through fusion of machine learning and fuzzy logic 一种融合机器学习和模糊逻辑的医学图像分类像素级解释卷积可解释性模型

Q2 Health Professions

Smart Health

Pub Date : 2024-12-21 DOI: 10.1016/j.smhl.2024.100535

Mohammad Ennab, Hamid Mcheick

Artificial intelligence (AI) models for medical image analysis have achieved high diagnostic performance, but they often lack interpretability, limiting their clinical adoption. Existing methods can explain predictions at the image level, but they cannot provide pixel-level insights. This study proposes a novel fusion of machine learning and fuzzy logic to develop an interpretable model that can precisely identify discriminative image regions driving diagnostic decisions and generate heatmap visualization. The model is trained and evaluated on a dataset of CT scans containing healthy and diseased organ images. Quantitative features are extracted across pixels and normalized into representation matrices using a machine learning model. Subsequently, the contribution of each detected lesion to the overall prediction is quantified using fuzzy logic. Organ segment weighted averages are computed to identify significant lesions. The model explains application of AI in medical imaging with an unprecedented level of detail. It can explain fine-grained image areas that have the greatest influence on diagnostic outcomes by mapping raw image pixels to fuzzy membership concepts. Lesions are found with effect sizes and statistical significance (p < 0.05).

Our model outperforms three existing methods in terms of interpretability and diagnostic accuracy by 10–15%, while maintaining computational efficiency. By disclosing crucial image evidence that supports AI decisions, this interpretable model improves transparency and clinician trust. Ethical implications of integrating AI in clinical settings are discussed, and future research directions are outlined. This study significantly advances the development of safe and interpretable AI for enhancing patient care through imaging analytics.

用于医学图像分析的人工智能（AI）模型已经取得了很高的诊断性能，但它们往往缺乏可解释性，限制了它们的临床应用。现有的方法可以在图像级别解释预测，但它们不能提供像素级别的洞察力。本研究提出了一种新的机器学习和模糊逻辑的融合，以开发一个可解释的模型，该模型可以精确识别驱动诊断决策的判别图像区域，并生成热图可视化。该模型在包含健康和病变器官图像的CT扫描数据集上进行训练和评估。通过机器学习模型提取像素间的定量特征，并将其归一化为表示矩阵。随后，使用模糊逻辑对每个检测到的病变对整体预测的贡献进行量化。计算器官节段加权平均值以识别重要病变。该模型以前所未有的详细程度解释了人工智能在医学成像中的应用。它可以通过将原始图像像素映射到模糊隶属度概念来解释对诊断结果影响最大的细粒度图像区域。发现病变具有效应量和统计学意义(p <；0.05)。我们的模型在保持计算效率的同时，在可解释性和诊断准确性方面优于现有的三种方法10-15%。通过披露支持人工智能决策的关键图像证据，这种可解释的模型提高了透明度和临床医生的信任。讨论了在临床环境中整合人工智能的伦理意义，并概述了未来的研究方向。这项研究极大地推动了安全、可解释的人工智能的发展，通过成像分析来加强患者护理。

{"title":"A novel convolutional interpretability model for pixel-level interpretation of medical image classification through fusion of machine learning and fuzzy logic","authors":"Mohammad Ennab, Hamid Mcheick","doi":"10.1016/j.smhl.2024.100535","DOIUrl":"10.1016/j.smhl.2024.100535","url":null,"abstract":"<div><div>Artificial intelligence (AI) models for medical image analysis have achieved high diagnostic performance, but they often lack interpretability, limiting their clinical adoption. Existing methods can explain predictions at the image level, but they cannot provide pixel-level insights. This study proposes a novel fusion of machine learning and fuzzy logic to develop an interpretable model that can precisely identify discriminative image regions driving diagnostic decisions and generate heatmap visualization. The model is trained and evaluated on a dataset of CT scans containing healthy and diseased organ images. Quantitative features are extracted across pixels and normalized into representation matrices using a machine learning model. Subsequently, the contribution of each detected lesion to the overall prediction is quantified using fuzzy logic. Organ segment weighted averages are computed to identify significant lesions. The model explains application of AI in medical imaging with an unprecedented level of detail. It can explain fine-grained image areas that have the greatest influence on diagnostic outcomes by mapping raw image pixels to fuzzy membership concepts. Lesions are found with effect sizes and statistical significance (p < 0.05).</div><div>Our model outperforms three existing methods in terms of interpretability and diagnostic accuracy by 10–15%, while maintaining computational efficiency. By disclosing crucial image evidence that supports AI decisions, this interpretable model improves transparency and clinician trust. Ethical implications of integrating AI in clinical settings are discussed, and future research directions are outlined. This study significantly advances the development of safe and interpretable AI for enhancing patient care through imaging analytics.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100535"},"PeriodicalIF":0.0,"publicationDate":"2024-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129217","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A novel guidance framework for nasal rapid antigen tests with improved swab keypoint detection 改进棉签关键点检测的鼻腔快速抗原检测新指导框架

Q2 Health Professions

Smart Health

Pub Date : 2024-12-06 DOI: 10.1016/j.smhl.2024.100534

Matthias Tschöpe, Dennis Schneider, Sungho Suh, Paul Lukowicz

The global impact of the COVID-19 pandemic has placed an unprecedented burden on healthcare systems. In this paper, we present a novel deep learning-based framework to guide individuals in performing nasal antigen rapid tests, with a particular focus on improving swab keypoint detection. Our system provides real-time feedback to participants on the correct execution of the test and may issue a certificate upon successful completion. While initially developed for COVID-19 antigen rapid tests, our versatile framework extends its applicability to various nasal screening tests, eliminating the need for specific information about the liquid solvent. To implement and evaluate our framework, we curated a comprehensive dataset with rapid test components and trained an object detection model to identify the position and size of all objects in each video frame. Addressing the challenge of swab depth classification, we propose a novel approach to locate and classify crucial swab points by a self-defined decision tree for depth assessment within the nasal cavity. The robustness of the proposed framework is validated with COVID-19 antigen rapid tests from various manufacturers. Experimental results demonstrate the remarkable performance of the framework in classifying the nasal placement of the swab, achieving an

F_{1}

-Score of 89.78%. Additionally, our framework attains an

F_{1}

-Score of 99.37% in classifying final test results on the test device.

COVID-19大流行的全球影响给卫生保健系统带来了前所未有的负担。在本文中，我们提出了一种新的基于深度学习的框架来指导个体进行鼻抗原快速检测，特别侧重于改进拭子关键点检测。我们的系统提供实时反馈给参与者正确执行测试，并可能在成功完成后颁发证书。虽然最初是为COVID-19抗原快速检测开发的，但我们的多功能框架将其适用性扩展到各种鼻筛查测试，从而消除了对液体溶剂特定信息的需求。为了实现和评估我们的框架，我们策划了一个包含快速测试组件的综合数据集，并训练了一个物体检测模型来识别每个视频帧中所有物体的位置和大小。为了解决拭子深度分类的难题，我们提出了一种新的方法，通过自定义的决策树来定位和分类鼻腔内深度评估的关键拭子点。通过不同制造商的COVID-19抗原快速检测，验证了所提出框架的稳健性。实验结果表明，该框架在棉签鼻腔放置分类方面表现出色，F1-Score达到89.78%。此外，我们的框架在对测试设备上的最终测试结果进行分类时达到了99.37%的F1-Score。

{"title":"A novel guidance framework for nasal rapid antigen tests with improved swab keypoint detection","authors":"Matthias Tschöpe, Dennis Schneider, Sungho Suh, Paul Lukowicz","doi":"10.1016/j.smhl.2024.100534","DOIUrl":"10.1016/j.smhl.2024.100534","url":null,"abstract":"<div><div>The global impact of the COVID-19 pandemic has placed an unprecedented burden on healthcare systems. In this paper, we present a novel deep learning-based framework to guide individuals in performing nasal antigen rapid tests, with a particular focus on improving swab keypoint detection. Our system provides real-time feedback to participants on the correct execution of the test and may issue a certificate upon successful completion. While initially developed for COVID-19 antigen rapid tests, our versatile framework extends its applicability to various nasal screening tests, eliminating the need for specific information about the liquid solvent. To implement and evaluate our framework, we curated a comprehensive dataset with rapid test components and trained an object detection model to identify the position and size of all objects in each video frame. Addressing the challenge of swab depth classification, we propose a novel approach to locate and classify crucial swab points by a self-defined decision tree for depth assessment within the nasal cavity. The robustness of the proposed framework is validated with COVID-19 antigen rapid tests from various manufacturers. Experimental results demonstrate the remarkable performance of the framework in classifying the nasal placement of the swab, achieving an <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span>-Score of 89.78%. Additionally, our framework attains an <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn></mrow></msub></math></span>-Score of 99.37% in classifying final test results on the test device.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100534"},"PeriodicalIF":0.0,"publicationDate":"2024-12-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129216","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Data-driven assessment of the effectiveness of non-pharmaceutical interventions on Covid spread mitigation in Italy 数据驱动评估意大利非药物干预措施对缓解Covid - 19传播的有效性

Q2 Health Professions

Smart Health

Pub Date : 2024-12-05 DOI: 10.1016/j.smhl.2024.100524

Divya Pragna Mulla , Mario Alessandro Bochicchio , Antonella Longo

To mitigate the impact of pandemics such as COVID-19, governments can implement various Non-Pharmaceutical Interventions (NPIs), ranging from the use of personal protective equipment to social distancing measures. While it has been demonstrated that NPIs can be effective over time, the assessment of their efficacy and the estimation of their cost-benefit ratio are still debated issues. For COVID-19, several authors have used case confirmation as a key parameter to assess the efficacy of NPIs. In this paper, we compare the efficacy of this parameter to that of the death rate, hospitalizations, and intensive care unit cases, in conjunction with human mobility indicators, in evaluating the effectiveness of NPIs. Our research uses data on daily COVID-19 cases and deaths, intensive care unit cases, hospitalizations, Google Mobility Reports, and NPI data from all Italian regions from March 2020 to May 2022. The evaluation method is based on the approach proposed by Wang et al., in 2020 to assess the impact of NPI efficacy and understand the effect of other parameters. Our results indicate that, when combined with human mobility indicators, the mortality rate and the number of intensive care units perform better than the number of cases in determining the efficacy of NPIs. These findings can assist policymakers in developing the best data-driven methods for dealing with confinement problems and planning for future outbreaks.

为了减轻COVID-19等大流行病的影响，各国政府可以实施各种非药物干预措施，从使用个人防护装备到保持社交距离措施。虽然已证明国家行动纲领可以长期有效，但评估其效力和估计其成本效益比仍然是有争议的问题。对于COVID-19，一些作者将病例确认作为评估npi效果的关键参数。在本文中，我们将该参数的有效性与死亡率、住院率和重症监护病房病例的有效性进行比较，并结合人类流动性指标来评估npi的有效性。我们的研究使用了2020年3月至2022年5月意大利所有地区的每日COVID-19病例和死亡、重症监护病房病例、住院情况、谷歌流动性报告和NPI数据。评估方法基于Wang等人在2020年提出的方法来评估NPI有效性的影响，并了解其他参数的影响。我们的研究结果表明，当与人的流动性指标相结合时，死亡率和重症监护病房数比病例数更能确定npi的疗效。这些发现可以帮助决策者制定最佳的数据驱动方法，以处理禁闭问题和规划未来的疫情。

{"title":"Data-driven assessment of the effectiveness of non-pharmaceutical interventions on Covid spread mitigation in Italy","authors":"Divya Pragna Mulla , Mario Alessandro Bochicchio , Antonella Longo","doi":"10.1016/j.smhl.2024.100524","DOIUrl":"10.1016/j.smhl.2024.100524","url":null,"abstract":"<div><div>To mitigate the impact of pandemics such as COVID-19, governments can implement various Non-Pharmaceutical Interventions (NPIs), ranging from the use of personal protective equipment to social distancing measures. While it has been demonstrated that NPIs can be effective over time, the assessment of their efficacy and the estimation of their cost-benefit ratio are still debated issues. For COVID-19, several authors have used case confirmation as a key parameter to assess the efficacy of NPIs. In this paper, we compare the efficacy of this parameter to that of the death rate, hospitalizations, and intensive care unit cases, in conjunction with human mobility indicators, in evaluating the effectiveness of NPIs. Our research uses data on daily COVID-19 cases and deaths, intensive care unit cases, hospitalizations, Google Mobility Reports, and NPI data from all Italian regions from March 2020 to May 2022. The evaluation method is based on the approach proposed by Wang et al., in 2020 to assess the impact of NPI efficacy and understand the effect of other parameters. Our results indicate that, when combined with human mobility indicators, the mortality rate and the number of intensive care units perform better than the number of cases in determining the efficacy of NPIs. These findings can assist policymakers in developing the best data-driven methods for dealing with confinement problems and planning for future outbreaks.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100524"},"PeriodicalIF":0.0,"publicationDate":"2024-12-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129220","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A novel rule-based expert system for early diagnosis of bipolar and Major Depressive Disorder 一种新的基于规则的双相情感障碍和重度抑郁症早期诊断专家系统

Q2 Health Professions

Smart Health

Pub Date : 2024-12-04 DOI: 10.1016/j.smhl.2024.100525

Mohammad Hossein Zolfagharnasab , Siavash Damari , Madjid Soltani , Artie Ng , Hengameh Karbalaeipour , Amin Haghdadi , Masood Hamed Saghayan , Farzam Matinfar

A confident and timely diagnosis of mental illnesses is one of the primary challenges practitioners repeatedly encounter when they start treating new patients. However, diagnosing can quickly become problematic as the subjects expose comparative symptoms among mental illnesses. Due to influencing a broad populace among mental ailments, an adjusted differentiation between Major Depressive Disorder, Mania Bipolar Disorder, Depressive Bipolar Disorder, and ordinary individuals with mild symptoms is one of the critical subjects for community health. This study responded to the described problem by proposing a novel rule-based Expert System, which evaluates the impact of disorder symptoms on the Certainty Factor concerning each mental status. The semantic rules are developed based on the recommendation of experts, and the implementation is carried out using Prolog and C# languages. Furthermore, an easy-to-use user interface is considered to facilitate the system workflow. The consistency of the developed framework is established by performing rigorous tests by expert psychiatrists as well as 120 clinical samples collected from private samples. Based on the results, the current model classifies mental disorder cases with a success rate of 93.33% using only the 17 symptoms specified in the ontology model. Furthermore, a questionnaire that measures user satisfaction after the test also achieves a mean score of 3.56 out of 4, which indicates a high degree of user acceptance. As a result, it is concluded that the current framework is a reliable tool for achieving a solid diagnosis in a shorter period.

自信而及时地诊断精神疾病是医生在开始治疗新病人时反复遇到的主要挑战之一。然而，诊断可能很快就会出现问题，因为受试者暴露了精神疾病之间的比较症状。重度抑郁症、躁狂性双相情感障碍、抑郁性双相情感障碍与症状轻微的普通个体之间的调整区分是社区卫生的重要课题之一。本研究通过提出一种新的基于规则的专家系统来回应所描述的问题，该系统评估障碍症状对涉及每种精神状态的确定性因素的影响。根据专家建议制定语义规则，并使用Prolog和c#语言进行实现。此外，一个易于使用的用户界面被认为是方便的系统工作流程。制定的框架的一致性是通过精神病专家进行严格测试以及从私人样本收集的120个临床样本来确定的。基于结果，目前的模型仅使用本体模型中指定的17种症状对精神障碍病例进行分类，成功率为93.33%。此外，测试后测量用户满意度的问卷也达到了3.56分（满分4分），这表明用户接受程度很高。因此，可以得出结论，目前的框架是在较短时间内实现可靠诊断的可靠工具。

{"title":"A novel rule-based expert system for early diagnosis of bipolar and Major Depressive Disorder","authors":"Mohammad Hossein Zolfagharnasab , Siavash Damari , Madjid Soltani , Artie Ng , Hengameh Karbalaeipour , Amin Haghdadi , Masood Hamed Saghayan , Farzam Matinfar","doi":"10.1016/j.smhl.2024.100525","DOIUrl":"10.1016/j.smhl.2024.100525","url":null,"abstract":"<div><div>A confident and timely diagnosis of mental illnesses is one of the primary challenges practitioners repeatedly encounter when they start treating new patients. However, diagnosing can quickly become problematic as the subjects expose comparative symptoms among mental illnesses. Due to influencing a broad populace among mental ailments, an adjusted differentiation between Major Depressive Disorder, Mania Bipolar Disorder, Depressive Bipolar Disorder, and ordinary individuals with mild symptoms is one of the critical subjects for community health. This study responded to the described problem by proposing a novel rule-based Expert System, which evaluates the impact of disorder symptoms on the Certainty Factor concerning each mental status. The semantic rules are developed based on the recommendation of experts, and the implementation is carried out using Prolog and <em>C#</em> languages. Furthermore, an easy-to-use user interface is considered to facilitate the system workflow. The consistency of the developed framework is established by performing rigorous tests by expert psychiatrists as well as 120 clinical samples collected from private samples. Based on the results, the current model classifies mental disorder cases with a success rate of 93.33% using only the 17 symptoms specified in the ontology model. Furthermore, a questionnaire that measures user satisfaction after the test also achieves a mean score of 3.56 out of 4, which indicates a high degree of user acceptance. As a result, it is concluded that the current framework is a reliable tool for achieving a solid diagnosis in a shorter period.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100525"},"PeriodicalIF":0.0,"publicationDate":"2024-12-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143129215","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Differences in gait parameters between supervised laboratory and unsupervised daily assessments of healthy adults measured with an in-shoe motion sensor system 用鞋内运动传感器系统测量有监督的实验室和无监督的健康成人每日评估之间的步态参数差异

Q2 Health Professions

Smart Health

Pub Date : 2024-11-26 DOI: 10.1016/j.smhl.2024.100526

Hiroki Shimizu , Takanobu Saito , Shione Kashiyama , Shinichi Kawamoto , Saori Morino , Momoko Nagai-Tanima , Tomoki Aoyama

This cross-sectional study compared the gait parameters between supervised laboratory and unsupervised daily life assessments in healthy adults. Gait was evaluated in 24 healthy young adults during 72 h of daily life and a 6-min laboratory gait at a comfortable speed. An in-shoe motion sensor system recorded gait data every 2 min, automatically detected stable gait segments by identifying repetitive movement patterns, and calculated the average of three consecutive valid gait cycles during each measurement period. Significant differences were found in walking speed (stride length divided by stride time; laboratory: 4.60 km/h vs. daily-life: 4.38 km/h), maximum (peak) dorsiflexion angle (laboratory: 29.71° vs. daily-life: 26.65°), maximum (peak) plantar flexion angle (laboratory: 74.54° vs. daily-life: 71.91°), roll angle of heel contact (laboratory: 7.46° vs. daily-life: 6.70°), maximum speed during the swing phase (laboratory: 14.49 km/h vs. daily-life: 12.68 km/h), circumduction (lateral displacement during the swing phase; laboratory: 2.68 cm vs. daily-life: 3.69 cm), toe-in/out angle (laboratory: 13.87° vs. daily-life: 15.32°), stance time (laboratory: 0.62 s vs. daily-life: 0.65 s), and pushing time (time between heel leaving and toe leaving the ground; laboratory: 0.20 s vs. daily-life: 0.21 s). The innovative aspect of this study is the comprehensive evaluation of foot-related gait parameters in real-world environments using an in-shoe motion sensor system. This approach provides ecologically valid insights into gait dynamics during daily activities, emphasizing the importance of real-world assessments for accurately evaluating gait performance and predicting adverse events such as falls. Keywords: Gait, Foot, Laboratory, Daily Life, Unsupervised Assessment, Shoe Motion Sensors.

这项横断面研究比较了有监督的实验室和无监督的健康成人日常生活评估的步态参数。对24名健康年轻人在日常生活72小时和以舒适速度进行6分钟的实验室步态进行评估。鞋内运动传感器系统每2分钟记录一次步态数据，通过识别重复运动模式自动检测稳定的步态片段，并计算每个测量周期内连续三个有效步态周期的平均值。行走速度(步幅除以步幅时间；实验室：4.60 km/h vs.日常：4.38 km/h)，最大（峰值）背屈角（实验室：29.71°vs.日常：26.65°），最大（峰值）足底屈角（实验室：74.54°vs.日常：71.91°），跟侧接触角（实验室：7.46°vs.日常：6.70°），摆动阶段的最大速度（实验室：14.49 km/h vs.日常：12.68 km/h），绕行(摆动阶段的侧向位移；实验室：2.68 cm vs.日常：3.69 cm)，脚趾进出角（实验室：13.87°vs.日常：15.32°），站立时间（实验室：0.62 s vs.日常：0.65 s），推蹬时间(脚跟离开到脚趾离开地面的时间；本研究的创新之处在于使用鞋内运动传感器系统对真实环境中与足部相关的步态参数进行综合评估。这种方法为日常活动中的步态动力学提供了生态学上有效的见解，强调了真实世界评估对准确评估步态性能和预测跌倒等不良事件的重要性。关键词：步态，足部，实验室，日常生活，无监督评估，鞋运动传感器。

{"title":"Differences in gait parameters between supervised laboratory and unsupervised daily assessments of healthy adults measured with an in-shoe motion sensor system","authors":"Hiroki Shimizu , Takanobu Saito , Shione Kashiyama , Shinichi Kawamoto , Saori Morino , Momoko Nagai-Tanima , Tomoki Aoyama","doi":"10.1016/j.smhl.2024.100526","DOIUrl":"10.1016/j.smhl.2024.100526","url":null,"abstract":"<div><div>This cross-sectional study compared the gait parameters between supervised laboratory and unsupervised daily life assessments in healthy adults. Gait was evaluated in 24 healthy young adults during 72 h of daily life and a 6-min laboratory gait at a comfortable speed. An in-shoe motion sensor system recorded gait data every 2 min, automatically detected stable gait segments by identifying repetitive movement patterns, and calculated the average of three consecutive valid gait cycles during each measurement period. Significant differences were found in walking speed (stride length divided by stride time; laboratory: 4.60 km/h vs. daily-life: 4.38 km/h), maximum (peak) dorsiflexion angle (laboratory: 29.71° vs. daily-life: 26.65°), maximum (peak) plantar flexion angle (laboratory: 74.54° vs. daily-life: 71.91°), roll angle of heel contact (laboratory: 7.46° vs. daily-life: 6.70°), maximum speed during the swing phase (laboratory: 14.49 km/h vs. daily-life: 12.68 km/h), circumduction (lateral displacement during the swing phase; laboratory: 2.68 cm vs. daily-life: 3.69 cm), toe-in/out angle (laboratory: 13.87° vs. daily-life: 15.32°), stance time (laboratory: 0.62 s vs. daily-life: 0.65 s), and pushing time (time between heel leaving and toe leaving the ground; laboratory: 0.20 s vs. daily-life: 0.21 s). The innovative aspect of this study is the comprehensive evaluation of foot-related gait parameters in real-world environments using an in-shoe motion sensor system. This approach provides ecologically valid insights into gait dynamics during daily activities, emphasizing the importance of real-world assessments for accurately evaluating gait performance and predicting adverse events such as falls. Keywords: Gait, Foot, Laboratory, Daily Life, Unsupervised Assessment, Shoe Motion Sensors.</div></div>","PeriodicalId":37151,"journal":{"name":"Smart Health","volume":"35 ","pages":"Article 100526"},"PeriodicalIF":0.0,"publicationDate":"2024-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142757705","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0