Ultrasonics. 2024 Mar 15;139:107293. doi: 10.1016/j.ultras.2024.107293. Online ahead of print.

ABSTRACT

We propose an innovative deep learning (DL) regression strategy combined with guided wave modes to address inverse acoustic scattering problems effectively. This approach allows for accurate recovery of heterogeneous defect fields at the interfaces of composite laminates. The neural network (NN) model's training process employs stochastic Gaussian fields as output, which are linked to the interfacial defect fields of the physical problem. Our method assumes prior knowledge of the material geometrical properties of the constituent layers. To model the interfaces, we utilize the Quasi-Static Approximation, a technique generating position-dependent interfacial stiffness matrices containing uncoupled normal and tangential springs. We validate our approach by assessing its performance in handling noisy input data and reduced models, as well as accounting model errors at the composite interface. The obtained results show that the proposed method has a remarkable generalization capability, allowing it to recover diverse defect field profiles with accuracy. Moreover, it exhibits robustness concerning noisy data and model errors. Lastly, thanks to the guided wave modes approach, the presented methodology not only maintains its capability to recover heterogeneous defect fields potentially in real-time but also extends the range of inspection to encompass a significantly larger structural area.

PMID:38492352 | DOI:10.1016/j.ultras.2024.107293

Radiol Phys Technol. 2024 Mar 16. doi: 10.1007/s12194-024-00794-x. Online ahead of print.

NO ABSTRACT

PMID:38492204 | DOI:10.1007/s12194-024-00794-x

Eur Radiol. 2024 Mar 16. doi: 10.1007/s00330-024-10693-9. Online ahead of print.

ABSTRACT

OBJECTIVE: To investigate whether a deep learning (DL) controlled aliasing in parallel imaging results in higher acceleration (CAIPIRINHA)-volumetric interpolated breath-hold examination (VIBE) technique can improve image quality, lesion conspicuity, and lesion detection compared to a standard CAIPIRINHA-VIBE technique in gadoxetic acid-enhanced liver MRI.

METHODS: This retrospective single-center study included 168 patients who underwent gadoxetic acid-enhanced liver MRI at 3 T using both standard CAIPIRINHA-VIBE and DL CAIPIRINHA-VIBE techniques on pre-contrast and hepatobiliary phase (HBP) images. Additionally, high-resolution (HR) DL CAIPIRINHA-VIBE was obtained with 1-mm slice thickness on the HBP. Three abdominal radiologists independently assessed the image quality and lesion conspicuity of pre-contrast and HBP images. Statistical analyses involved the Wilcoxon signed-rank test for image quality assessment and the generalized estimation equation for lesion conspicuity and detection evaluation.

RESULTS: DL and HR-DL CAIPIRINHA-VIBE demonstrated significantly improved overall image quality and reduced artifacts on pre-contrast and HBP images compared to standard CAIPIRINHA-VIBE (p < 0.001), with a shorter acquisition time (DL vs standard, 11 s vs 17 s). However, the former presented a more synthetic appearance (both p < 0.05). HR-DL CAIPIRINHA-VIBE showed superior lesion conspicuity to standard and DL CAIPIRINHA-VIBE on HBP images (p < 0.001). Moreover, HR-DL CAIPIRINHA-VIBE exhibited a significantly higher detection rate of small (< 2 cm) solid focal liver lesions (FLLs) on HBP images compared to standard CAIPIRINHA-VIBE (92.5% vs 87.4%; odds ratio = 1.83; p = 0.036).

CONCLUSION: DL and HR-DL CAIPIRINHA-VIBE achieved superior image quality compared to standard CAIPIRINHA-VIBE. Additionally, HR-DL CAIPIRINHA-VIBE improved the lesion conspicuity and detection of small solid FLLs. DL and HR-DL CAIPIRINHA-VIBE hold the potential clinical utility for gadoxetic acid-enhanced liver MRI.

CLINICAL RELEVANCE STATEMENT: DL and HR-DL CAIPIRINHA-VIBE hold promise as potential alternatives to standard CAIPIRINHA-VIBE in routine clinical liver MRI, improving the image quality and lesion conspicuity, enhancing the detection of small (< 2 cm) solid focal liver lesions, and reducing the acquisition time.

KEY POINTS: • DL and HR-DL CAIPIRINHA-VIBE demonstrated improved overall image quality and reduced artifacts on pre-contrast and HBP images compared to standard CAIPIRINHA-VIBE, in addition to a shorter acquisition time. • DL and HR-DL CAIPIRINHA-VIBE yielded a more synthetic appearance than standard CAIPIRINHA-VIBE. • HR-DL CAIPIRINHA-VIBE showed improved lesion conspicuity than standard CAIPIRINHA-VIBE on HBP images, with a higher detection of small (< 2 cm) solid focal liver lesions.

PMID:38492004 | DOI:10.1007/s00330-024-10693-9

Chem Biol Drug Des. 2024 Mar;103(3):e14505. doi: 10.1111/cbdd.14505.

ABSTRACT

Human beings possess trillions of microbial cells in a symbiotic relationship. This relationship benefits both partners for a long time. The gut microbiota helps in many bodily functions from harvesting energy from digested food to strengthening biochemical barriers of the gut and intestine. But the changes in microbiota composition and bacteria that can enter the gastrointestinal tract can cause infection. Several approaches like culture-independent techniques such as high-throughput and meta-omics projects targeting 16S ribosomal RNA (rRNA) sequencing are popular methods to investigate the composition of the human gastrointestinal tract microbiota and taxonomically characterizing microbial communities. The microbiota conformation and diversity should be provided by whole-genome shotgun metagenomic sequencing of site-specific community DNA associating genome mapping, gene inventory, and metabolic remodelling and reformation, to ease the functional study of human microbiota. Preliminary examination of the therapeutic potency for dysbiosis-associated diseases permits investigation of pharmacokinetic-pharmacodynamic changes in microbial communities for escalation of treatment and dosage plan. Gut microbiome study is an integration of metagenomics which has influenced the field in the last two decades. And the incorporation of artificial intelligence and deep learning through "omics-based" methods and microfluidic evaluation enhanced the capability of identification of thousands of microbes.

PMID:38491814 | DOI:10.1111/cbdd.14505

NMR Biomed. 2024 Mar 15:e5130. doi: 10.1002/nbm.5130. Online ahead of print.

ABSTRACT

Chemical exchange saturation transfer (CEST) MRI is a molecular imaging tool that provides physiological information about tissues, making it an invaluable tool for disease diagnosis and guided treatment. Its clinical application requires the acquisition of high-resolution images capable of accurately identifying subtle regional changes in vivo, while simultaneously maintaining a high level of spectral resolution. However, the acquisition of such high-resolution images is time consuming, presenting a challenge for practical implementation in clinical settings. Among several techniques that have been explored to reduce the acquisition time in MRI, deep-learning-based super-resolution (DLSR) is a promising approach to address this problem due to its adaptability to any acquisition sequence and hardware. However, its translation to CEST MRI has been hindered by the lack of the large CEST datasets required for network development. Thus, we aim to develop a DLSR method, named DLSR-CEST, to reduce the acquisition time for CEST MRI by reconstructing high-resolution images from fast low-resolution acquisitions. This is achieved by first pretraining the DLSR-CEST on human brain T1w and T2w images to initialize the weights of the network and then training the network on very small human and mouse brain CEST datasets to fine-tune the weights. Using the trained DLSR-CEST network, the reconstructed CEST source images exhibited improved spatial resolution in both peak signal-to-noise ratio and structural similarity index measure metrics at all downsampling factors (2-8). Moreover, amide CEST and relayed nuclear Overhauser effect maps extrapolated from the DLSR-CEST source images exhibited high spatial resolution and low normalized root mean square error, indicating a negligible loss in Z-spectrum information. Therefore, our DLSR-CEST demonstrated a robust reconstruction of high-resolution CEST source images from fast low-resolution acquisitions, thereby improving the spatial resolution and preserving most Z-spectrum information.

PMID:38491754 | DOI:10.1002/nbm.5130

Biomed Tech (Berl). 2024 Mar 18. doi: 10.1515/bmt-2023-0126. Online ahead of print.

ABSTRACT

OBJECTIVES: In this study, we developed a machine learning approach for postoperative corneal endothelial cell images of patients who underwent Descemet's membrane keratoplasty (DMEK).

METHODS: An AlexNet model is proposed and validated throughout the study for endothelial cell segmentation and cell location determination. The 506 images of postoperative corneal endothelial cells were analyzed. Endothelial cell detection, segmentation, and determining of its polygonal structure were identified. The proposed model is based on the training of an R-CNN to locate endothelial cells. Next, by determining the ridges separating adjacent cells, the density and hexagonality rates of DMEK patients are calculated.

RESULTS: The proposed method reached accuracy and F1 score rates of 86.15 % and 0.857, respectively, which indicates that it can reliably replace the manual detection of cells in vivo confocal microscopy (IVCM). The AUC score of 0.764 from the proposed segmentation method suggests a satisfactory outcome.

CONCLUSIONS: A model focused on segmenting endothelial cells can be employed to assess the health of the endothelium in DMEK patients.

PMID:38491745 | DOI:10.1515/bmt-2023-0126

Biomed Eng Online. 2024 Mar 15;23(1):34. doi: 10.1186/s12938-024-01227-x.

ABSTRACT

BACKGROUND: Decubitus ulcers are prevalent among the aging population due to a gradual decline in their overall health, such as nutrition, mental health, and mobility, resulting in injury to the skin and tissue. The most common technique to prevent these ulcers is through frequent repositioning to redistribute body pressures. Therefore, the main goal of this study is to facilitate the timely repositioning of patients through the use of a pressure mat to identify in-bed postures in various sleep environments. Pressure data were collected from 10 healthy participants lying down on a pressure mat in 19 various in-bed postures, correlating to the supine, prone, right-side, and left-side classes. In addition, pressure data were collected from participants sitting at the edge of the bed as well as an empty bed. Each participant was asked to lie in these 19 postures in three distinct testing environments: a hospital bed, a home bed, and a home bed with a foam mattress topper. To categorize each posture into its respective class, the pre-trained 2D ResNet-18 CNN and the pre-trained Inflated 3D CNN algorithms were trained and validated using image and video pressure mapped data, respectively.

RESULTS: The ResNet-18 and Inflated 3D CNN algorithms were validated using leave-one-subject-out (LOSO) and leave-one-environment-out (LOEO) cross-validation techniques. LOSO provided an average accuracy of 92.07% ± 5.72% and 82.22% ± 8.50%, for the ResNet-18 and Inflated 3D CNN algorithms, respectively. Contrastingly, LOEO provided a reduced average accuracy of 85.37% ± 14.38% and 77.79% ± 9.76%, for the ResNet-18 and Inflated 3D CNN algorithms, respectively.

CONCLUSION: These pilot results indicate that the proposed algorithms can accurately distinguish between in-bed postures, on unseen participant data as well as unseen mattress environment data. The proposed algorithms can establish the basis of a decubitus ulcer prevention platform that can be applied to various sleeping environments. To the best of our knowledge, the impact of mattress stiffness has not been considered in previous studies regarding in-bed posture monitoring.

PMID:38491463 | DOI:10.1186/s12938-024-01227-x

BMC Ophthalmol. 2024 Mar 15;24(1):121. doi: 10.1186/s12886-024-03381-1.

ABSTRACT

Age related macular degeneration (AMD) represents a leading cause of vision loss and it is expected to affect 288 million people by 2040. During the last decade, machine learning technologies have shown great potential to revolutionize clinical management of AMD and support research for a better understanding of the disease. The aim of this review is to provide a panoramic description of all the applications of AI to AMD management and screening that have been analyzed in recent past literature. Deep learning (DL) can be effectively used to diagnose AMD, to predict short term risk of exudation and need for injections within the next 2 years. Moreover, DL technology has the potential to customize anti-VEGF treatment choice with a higher accuracy than expert human experts. In addition, accurate prediction of VA response to treatment can be provided to the patients with the use of ML models, which could considerably increase patients' compliance to treatment in favorable cases. Lastly, AI, especially in the form of DL, can effectively predict conversion to GA in 12 months and also suggest new biomarkers of conversion with an innovative reverse engineering approach.

PMID:38491380 | DOI:10.1186/s12886-024-03381-1

Int J Comput Assist Radiol Surg. 2024 Mar 15. doi: 10.1007/s11548-024-03089-z. Online ahead of print.

ABSTRACT

PURPOSE: Pelvic X-ray (PXR) is widely utilized in clinical decision-making associated with the pelvis, the lower part of the trunk that supports and balances the trunk. In particular, PXR-based landmark detection facilitates downstream analysis and computer-assisted diagnosis and treatment of pelvic diseases. Although PXR has the advantages of low radiation and reduced cost compared to computed tomography (CT), it characterizes the 2D pelvis-tissue superposition of 3D structures, which may affect the accuracy of landmark detection in some cases. However, the superposition nature of PXR is implicitly handled by existing deep learning-based landmark detection methods, which mainly design the deep network structures for better detection performances. Explicit handling of the superposition nature of PXR is rarely done.

METHODS: In this paper, we explicitly focus on the superposition of X-ray images. Specifically, we propose a pelvis extraction (PELE) module that consists of a decomposition network, a domain adaptation network, and an enhancement module, which utilizes 3D prior anatomical knowledge in CT to guide and well isolate the pelvis from PXR, thereby eliminating the influence of soft tissue for landmark detection. The extracted pelvis image, after enhancement, is then used for landmark detection.

RESULTS: We conduct an extensive evaluation based on two public and one private dataset, totaling 850 PXRs. The experimental results show that the proposed PELE module significantly improves the accuracy of PXRs landmark detection and achieves state-of-the-art performances in several benchmark metrics.

CONCLUSION: The design of PELE module can improve the accuracy of different pelvic landmark detection baselines, which we believe is obviously conducive to the positioning and inspection of clinical landmarks and critical structures, thus better serving downstream tasks. Our project has been open-sourced at https://github.com/ECNUACRush/PELEscores .

PMID:38491244 | DOI:10.1007/s11548-024-03089-z

J Imaging Inform Med. 2024 Mar 15. doi: 10.1007/s10278-024-01062-5. Online ahead of print.

ABSTRACT

In this article, we propose an AI-based low-risk visualization framework for lung health monitoring using low-resolution ultra-low-dose CT (LR-ULDCT). We present a novel deep cascade processing workflow to achieve diagnostic visualization on LR-ULDCT (<0.3 mSv) at par high-resolution CT (HRCT) of 100 mSV radiation technology. To this end, we build a low-risk and affordable deep cascade network comprising three sequential deep processes: restoration, super-resolution (SR), and segmentation. Given degraded LR-ULDCT, the first novel network unsupervisedly learns restoration function from augmenting patch-based dictionaries and residuals. The restored version is then super-resolved (SR) for target (sensor) resolution. Here, we combine perceptual and adversarial losses in novel GAN to establish the closeness between probability distributions of generated SR-ULDCT and restored LR-ULDCT. Thus SR-ULDCT is presented to the segmentation network that first separates the chest portion from SR-ULDCT followed by lobe-wise colorization. Finally, we extract five lobes to account for the presence of ground glass opacity (GGO) in the lung. Hence, our AI-based system provides low-risk visualization of input degraded LR-ULDCT to various stages, i.e., restored LR-ULDCT, restored SR-ULDCT, and segmented SR-ULDCT, and achieves diagnostic power of HRCT. We perform case studies by experimenting on real datasets of COVID-19, pneumonia, and pulmonary edema/congestion while comparing our results with state-of-the-art. Ablation experiments are conducted for better visualizing different operating pipelines. Finally, we present a verification report by fourteen (14) experienced radiologists and pulmonologists.

PMID:38491236 | DOI:10.1007/s10278-024-01062-5

J Imaging Inform Med. 2024 Mar 15. doi: 10.1007/s10278-024-01067-0. Online ahead of print.

ABSTRACT

Our study aims to evaluate the potential of a deep learning (DL) algorithm for differentiating the signal intensity of bone marrow between osteomyelitis (OM), Charcot neuropathic osteoarthropathy (CNO), and trauma (TR). The local ethics committee approved this retrospective study. From 148 patients, segmentation resulted in 679 labeled regions for T1-weighted images (comprising 151 CNO, 257 OM, and 271 TR) and 714 labeled regions for T2-weighted images (consisting of 160 CNO, 272 OM, and 282 TR). We employed both multi-class classification (MCC) and binary-class classification (BCC) approaches to compare the classification outcomes of CNO, TR, and OM. The ResNet-50 and the EfficientNet-b0 accuracy values were computed at 96.2% and 97.1%, respectively, for T1-weighted images. Additionally, accuracy values for ResNet-50 and the EfficientNet-b0 were determined at 95.6% and 96.8%, respectively, for T2-weighted images. Also, according to BCC for CNO, OM, and TR, the sensitivity of ResNet-50 is 91.1%, 92.4%, and 96.6% and the sensitivity of EfficientNet-b0 is 93.2%, 97.6%, and 98.1% for T1, respectively. For CNO, OM, and TR, the sensitivity of ResNet-50 is 94.9%, 83.6%, and 97.9% and the sensitivity of EfficientNet-b0 is 95.6%, 85.2%, and 98.6% for T2, respectively. The specificity values of ResNet-50 for CNO, OM, and TR in T1-weighted images are 98.1%, 97.9%, and 94.7% and 98.6%, 97.5%, and 96.7% in T2-weighted images respectively. Similarly, for EfficientNet-b0, the specificity values are 98.9%, 98.7%, and 98.4% and 99.1%, 98.5%, and 98.7% for T1-weighted and T2-weighted images respectively. In the diabetic foot, deep learning methods serve as a non-invasive tool to differentiate CNO, OM, and TR with high accuracy.

PMID:38491234 | DOI:10.1007/s10278-024-01067-0

Sci Prog. 2024 Jan-Mar;107(1):368504241236557. doi: 10.1177/00368504241236557.

ABSTRACT

We introduce a comprehensive analysis of several approaches used in stock price forecasting, including statistical, machine learning, and deep learning models. The advantages and limitations of these models are discussed to provide an insight into stock price forecasting. Traditional statistical methods, such as the autoregressive integrated moving average and its variants, are recognized for their efficiency, but they also have some limitations in addressing non-linear problems and providing long-term forecasts. Machine learning approaches, including algorithms such as artificial neural networks and random forests, are praised for their ability to grasp non-linear information without depending on stochastic data or economic theory. Moreover, deep learning approaches, such as convolutional neural networks and recurrent neural networks, can deal with complex patterns in stock prices. Additionally, this study further investigates hybrid models, combining various approaches to explore their strengths and counterbalance individual weaknesses, thereby enhancing predictive accuracy. By presenting a detailed review of various studies and methods, this study illuminates the direction of stock price forecasting and highlights potential approaches for further studies refining the stock price forecasting models.

PMID:38490223 | DOI:10.1177/00368504241236557

Sci Prog. 2024 Jan-Mar;107(1):368504241236345. doi: 10.1177/00368504241236345.

ABSTRACT

The accurate identification of dynamic change of limb length discrepancy (LLD) in non-clinical settings is of great significance for monitoring gait function change in people's everyday lives. How to search for advanced techniques to measure LLD changes in non-clinical settings has always been a challenging endeavor in recent related research. In this study, we have proposed a novel approach to accurately measure the dynamic change of LLD outdoors by using deep learning and wearable sensors. The basic idea is that the measurement of dynamic change of LLD was considered as a multiple gait classification task based on LLD change that is clearly associated with its gait pattern. A hybrid deep learning model of convolutional neural network and long short-term memory (CNN-LSTM) was developed to precisely classify LLD gait patterns by discovering the most representative spatial-temporal LLD dynamic change features. Twenty-three healthy subjects were recruited to simulate four levels of LLD by wearing a shoe lift with different heights. The Delsys TrignoTM system was implemented to simultaneously acquire gait data from six sensors positioned on the hip, knee and ankle joint of two lower limbs respectively. The experimental results showed that the developed CNN-LSTM model could reach a higher accuracy of 93.24% and F1-score of 93.48% to classify four different LLD gait patterns when compared with CNN, LSTM, and CNN-gated recurrent unit(CNN-GRU), and gain better recall and precision (more than 92%) to detect each LLD gait pattern accurately. Our model could achieve excellent learning ability to discover the most representative LLD dynamic change features for classifying LLD gait patterns accurately. Our technical solution would help not only to accurately measure LLD dynamic change in non-clinical settings, but also to potentially find out lower limb joints with more abnormal compensatory change caused by LLD.

PMID:38490169 | DOI:10.1177/00368504241236345

Spectrochim Acta A Mol Biomol Spectrosc. 2024 Mar 12;313:124147. doi: 10.1016/j.saa.2024.124147. Online ahead of print.

ABSTRACT

The accurate and effective detection of moisture in coal gravels is crucial. Conventional air oven-drying method suffers from prolonged processing times and their disruptive nature. This paper proposes a deep learning-based adaptive fusion method for multiple microwave spectra to non-destructively detect the moisture content of coal gravels. First, a purpose-built free-space measurement platform is employed to acquire microwave spectra of coal samples, encompassing the magnitude and phase spectra of reflection coefficients (S11) and transmission coefficients (S21). Subsequently, a Monte-Carlo cross-validation-based method is adopted to detect and eliminate outliers in the spectra. Furthermore, a novel feature extraction module is proposed, enhancing the traditional U-shaped network using residual learning (ResNet) and the convolutional block attention module (CBAM) to extract and reconstruct subtle spectral features. Inspired by the high-level data fusion, an adaptive spectra fusion method is then introduced that can autonomously balance the contributions between different spectra. The experimental results underscore the advantages of the proposed method, with narrow frequency intervals between 2.50-3.25 GHz, 3.75-4.00 GHz, and 4.75-5.00 GHz exhibiting superior detection accuracy compared to the entire frequency band, achieving R2 = 0.9034, MAE = 1.0254, RMSE = 1.2948 and RPIQ = 6.0630.

PMID:38490123 | DOI:10.1016/j.saa.2024.124147

Int J Neural Syst. 2024 Mar 16:2450026. doi: 10.1142/S0129065724500266. Online ahead of print.

ABSTRACT

The global trend of increasing life expectancy introduces new challenges with far-reaching implications. Among these, the risk of falls among older adults is particularly significant, affecting individual health and the quality of life, and placing an additional burden on healthcare systems. Existing fall detection systems often have limitations, including delays due to continuous server communication, high false-positive rates, low adoption rates due to wearability and comfort issues, and high costs. In response to these challenges, this work presents a reliable, wearable, and cost-effective fall detection system. The proposed system consists of a fit-for-purpose device, with an embedded algorithm and an Inertial Measurement Unit (IMU), enabling real-time fall detection. The algorithm combines a Threshold-Based Algorithm (TBA) and a neural network with low number of parameters based on a Transformer architecture. This system demonstrates notable performance with 95.29% accuracy, 93.68% specificity, and 96.66% sensitivity, while only using a 0.38% of the trainable parameters used by the other approach.

PMID:38490957 | DOI:10.1142/S0129065724500266

Biochim Biophys Acta Gen Subj. 2024 Mar 13:130597. doi: 10.1016/j.bbagen.2024.130597. Online ahead of print.

ABSTRACT

BACKGROUND: Abiotic stresses pose serious threat to the growth and yield of crop plants. Several studies suggest that in plants, transcription factors (TFs) are important regulators of gene expression, especially when it comes to coping with abiotic stresses. Therefore, it is crucial to identify TFs associated with abiotic stress response for breeding of abiotic stress tolerant crop cultivars.

METHODS: Based on a machine learning frame work, a computational model was envisaged to predict TFs associated with abiotic stress response in plants. To numerically encode TF sequences, four distinct sequence derived features were generated. The prediction was performed using ten shallow learning and four deep learning algorithms. For prediction using more pertinent and informative features, feature selection techniques were also employed.

RESULTS: Using the features chosen by the light-gradient boosting machine-variable importance measure (LGBM-VIM), the LGBM achieved the highest cross-validation performance metrics (accuracy: 86.81%, auROC: 92.98%, and auPRC: 94.03%). Further evaluation of the proposed model (LGBM prediction method + LGBM-VIM selected features) was also done using an independent test dataset, where the accuracy, auROC and auPRC were observed 81.98%, 90.65% and 91.30%, respectively.

CONCLUSIONS: To facilitate the adoption of the proposed strategy by users, the technique was implemented and made available as a prediction server called ASPTF, accessible at https://iasri-sg.icar.gov.in/asptf/. The developed approach and the corresponding web application are anticipated to complement experimental methods in the identification of transcription factors (TFs) responsive to abiotic stress in plants.

PMID:38490467 | DOI:10.1016/j.bbagen.2024.130597

Neural Netw. 2024 Mar 5;174:106226. doi: 10.1016/j.neunet.2024.106226. Online ahead of print.

ABSTRACT

Convolutional neural networks (CNNs) have gained immense popularity in recent years, finding their utility in diverse fields such as image recognition, natural language processing, and bio-informatics. Despite the remarkable progress made in deep learning theory, most studies on CNNs, especially in regression tasks, tend to heavily rely on the least squares loss function. However, there are situations where such learning algorithms may not suffice, particularly in the presence of heavy-tailed noises or outliers. This predicament emphasizes the necessity of exploring alternative loss functions that can handle such scenarios more effectively, thereby unleashing the true potential of CNNs. In this paper, we investigate the generalization error of deep CNNs with the rectified linear unit (ReLU) activation function for robust regression problems within an information-theoretic learning framework. Our study demonstrates that when the regression function exhibits an additive ridge structure and the noise possesses a finite pth moment, the empirical risk minimization scheme, generated by the maximum correntropy criterion and deep CNNs, achieves fast convergence rates. Notably, these rates align with the mini-max optimal convergence rates attained by fully connected neural network model with the Huber loss function up to a logarithmic factor. Additionally, we further establish the convergence rates of deep CNNs under the maximum correntropy criterion when the regression function resides in a Sobolev space on the sphere.

PMID:38490117 | DOI:10.1016/j.neunet.2024.106226

Neural Netw. 2024 Mar 11;174:106230. doi: 10.1016/j.neunet.2024.106230. Online ahead of print.

ABSTRACT

Unsupervised domain adaptation (UDA) via deep learning has attracted appealing attention for tackling domain-shift problems caused by distribution discrepancy across different domains. Existing UDA approaches highly depend on the accessibility of source domain data, which is usually limited in practical scenarios due to privacy protection, data storage and transmission cost, and computation burden. To tackle this issue, many source-free unsupervised domain adaptation (SFUDA) methods have been proposed recently, which perform knowledge transfer from a pre-trained source model to the unlabeled target domain with source data inaccessible. A comprehensive review of these works on SFUDA is of great significance. In this paper, we provide a timely and systematic literature review of existing SFUDA approaches from a technical perspective. Specifically, we categorize current SFUDA studies into two groups, i.e., white-box SFUDA and black-box SFUDA, and further divide them into finer subcategories based on different learning strategies they use. We also investigate the challenges of methods in each subcategory, discuss the advantages/disadvantages of white-box and black-box SFUDA methods, conclude the commonly used benchmark datasets, and summarize the popular techniques for improved generalizability of models learned without using source data. We finally discuss several promising future directions in this field.

PMID:38490115 | DOI:10.1016/j.neunet.2024.106230

Comput Biol Med. 2024 Feb 28;172:108207. doi: 10.1016/j.compbiomed.2024.108207. Online ahead of print.

ABSTRACT

Artificial Intelligence (AI) techniques are increasingly used in computer-aided diagnostic tools in medicine. These techniques can also help to identify Hypertension (HTN) in its early stage, as it is a global health issue. Automated HTN detection uses socio-demographic, clinical data, and physiological signals. Additionally, signs of secondary HTN can also be identified using various imaging modalities. This systematic review examines related work on automated HTN detection. We identify datasets, techniques, and classifiers used to develop AI models from clinical data, physiological signals, and fused data (a combination of both). Image-based models for assessing secondary HTN are also reviewed. The majority of the studies have primarily utilized single-modality approaches, such as biological signals (e.g., electrocardiography, photoplethysmography), and medical imaging (e.g., magnetic resonance angiography, ultrasound). Surprisingly, only a small portion of the studies (22 out of 122) utilized a multi-modal fusion approach combining data from different sources. Even fewer investigated integrating clinical data, physiological signals, and medical imaging to understand the intricate relationships between these factors. Future research directions are discussed that could build better healthcare systems for early HTN detection through more integrated modeling of multi-modal data sources.

PMID:38489986 | DOI:10.1016/j.compbiomed.2024.108207

Comput Methods Programs Biomed. 2024 Mar 12;248:108118. doi: 10.1016/j.cmpb.2024.108118. Online ahead of print.

ABSTRACT

BACKGROUND: Estimating the risk of a difficult tracheal intubation should help clinicians in better anaesthesia planning, to maximize patient safety. Routine bedside screenings suffer from low sensitivity.

OBJECTIVE: To develop and evaluate machine learning (ML) and deep learning (DL) algorithms for the reliable prediction of intubation risk, using information about airway morphology.

METHODS: Observational, prospective cohort study enrolling n=623 patients who underwent tracheal intubation: 53/623 difficult cases (prevalence 8.51%). First, we used our previously validated deep convolutional neural network (DCNN) to extract 2D image coordinates for 27 + 13 relevant anatomical landmarks in two preoperative photos (frontal and lateral views). Here we propose a method to determine the 3D pose of the camera with respect to the patient and to obtain the 3D world coordinates of these landmarks. Then we compute a novel set of dM=59 morphological features (distances, areas, angles and ratios), engineered with our anaesthesiologists to characterize each individual's airway anatomy towards prediction. Subsequently, here we propose four ad hoc ML pipelines for difficult intubation prognosis, each with four stages: feature scaling, imputation, resampling for imbalanced learning, and binary classification (Logistic Regression, Support Vector Machines, Random Forests and eXtreme Gradient Boosting). These compound ML pipelines were fed with the dM=59 morphological features, alongside dD=7 demographic variables. Here we trained them with automatic hyperparameter tuning (Bayesian search) and probability calibration (Platt scaling). In addition, we developed an ad hoc multi-input DCNN to estimate the intubation risk directly from each pair of photographs, i.e. without any intermediate morphological description. Performance was evaluated using optimal Bayesian decision theory. It was compared against experts' judgement and against state-of-the-art methods (three clinical formulae, four ML, four DL models).

RESULTS: Our four ad hoc ML pipelines with engineered morphological features achieved similar discrimination capabilities: median AUCs between 0.746 and 0.766. They significantly outperformed both expert judgement and all state-of-the-art methods (highest AUC at 0.716). Conversely, our multi-input DCNN yielded low performance due to overfitting. This same behaviour occurred for the state-of-the-art DL algorithms. Overall, the best method was our XGB pipeline, with the fewest false negatives at the optimal Bayesian decision threshold.

CONCLUSIONS: We proposed and validated ML models to assist clinicians in anaesthesia planning, providing a reliable calibrated estimate of airway intubation risk, which outperformed expert assessments and state-of-the-art methods. Our novel set of engineered features succeeded in providing informative descriptions for prognosis.

PMID:38489935 | DOI:10.1016/j.cmpb.2024.108118