Front Plant Sci. 2024 Apr 10;15:1375118. doi: 10.3389/fpls.2024.1375118. eCollection 2024.

ABSTRACT

In order to address the challenges of inefficiency and insufficient accuracy in the manual identification of young citrus fruits during thinning processes, this study proposes a detection methodology using the you only look once for complex backgrounds of young citrus fruits (YCCB-YOLO) approach. The method first constructs a dataset containing images of young citrus fruits in a real orchard environment. To improve the detection accuracy while maintaining the computational efficiency, the study reconstructs the detection head and backbone network using pointwise convolution (PWonv) lightweight network, which reduces the complexity of the model without affecting the performance. In addition, the ability of the model to accurately detect young citrus fruits in complex backgrounds is enhanced by integrating the fusion attention mechanism. Meanwhile, the simplified spatial pyramid pooling fast-large kernel separated attention (SimSPPF-LSKA) feature pyramid was introduced to further enhance the multi-feature extraction capability of the model. Finally, the Adam optimization function was used to strengthen the nonlinear representation and feature extraction ability of the model. The experimental results show that the model achieves 91.79% precision (P), 92.75% recall (R), and 97.32% mean average precision (mAP)on the test set, which were improved by 1.33%, 2.24%, and 1.73%, respectively, compared with the original model, and the size of the model is only 5.4 MB. This study could meet the performance requirements for citrus fruit identification, which provides technical support for fruit thinning.

PMID:38660450 | PMC:PMC11039839 | DOI:10.3389/fpls.2024.1375118

Front Public Health. 2024 Apr 10;12:1386110. doi: 10.3389/fpubh.2024.1386110. eCollection 2024.

ABSTRACT

PURPOSE: Artificial intelligence has led to significant developments in the healthcare sector, as in other sectors and fields. In light of its significance, the present study delves into exploring deep learning, a branch of artificial intelligence.

METHODS: In the study, deep learning networks ResNet101, AlexNet, GoogLeNet, and Xception were considered, and it was aimed to determine the success of these networks in disease diagnosis. For this purpose, a dataset of 1,680 chest X-ray images was utilized, consisting of cases of COVID-19, viral pneumonia, and individuals without these diseases. These images were obtained by employing a rotation method to generate replicated data, wherein a split of 70 and 30% was adopted for training and validation, respectively.

RESULTS: The analysis findings revealed that the deep learning networks were successful in classifying COVID-19, Viral Pneumonia, and Normal (disease-free) images. Moreover, an examination of the success levels revealed that the ResNet101 deep learning network was more successful than the others with a 96.32% success rate.

CONCLUSION: In the study, it was seen that deep learning can be used in disease diagnosis and can help experts in the relevant field, ultimately contributing to healthcare organizations and the practices of country managers.

PMID:38660365 | PMC:PMC11039909 | DOI:10.3389/fpubh.2024.1386110

Heliyon. 2024 Apr 16;10(8):e29677. doi: 10.1016/j.heliyon.2024.e29677. eCollection 2024 Apr 30.

ABSTRACT

Pelvic malalignment leads to general imbalance and adversely affects leg length. Timely and accurate diagnosis of pelvic alignment in patients is crucial to prevent additional complications arising from delayed treatment. Currently, doctors typically assess pelvic alignment either manually or through radiography. This study aimed to develop and assess the validity of a deep learning-based system for automatically measuring 10 radiographic parameters necessary for diagnosing pelvic displacement using standing anteroposterior pelvic X-rays. Between March 2016 and June 2021, pelvic radiographs from 1215 patients were collected. After applying specific selection criteria, 550 pelvic radiographs were chosen for analysis. These data were utilized to develop a deep learning-based system capable of automatically measuring radiographic parameters relevant to pelvic displacement diagnosis. The system's diagnostic accuracy was evaluated by comparing automatically measured values with those assessed by a clinician using 200 radiographs selected from the initial 550. The results indicated that the system exhibited high reliability, accuracy, and reproducibility, with a Pearson correlation coefficient of ≥0.9, an intra-class correlation coefficient of ≥0.9, a mean absolute error of ≤1 cm, mean square error of ≤1 cm, and root mean square error of ≤1 cm. Moreover, the system's measurement time for a single radiograph was found to be 18 to 20 times faster than that required by a clinician for manual inspection. In conclusion, our proposed deep learning-based system effectively utilizes standing anteroposterior pelvic radiographs to precisely and consistently measure radiographic parameters essential for diagnosing pelvic displacement.

PMID:38660256 | PMC:PMC11040132 | DOI:10.1016/j.heliyon.2024.e29677

Front Neurosci. 2024 Apr 10;18:1329718. doi: 10.3389/fnins.2024.1329718. eCollection 2024.

ABSTRACT

PURPOSE: To develop deep learning models based on four-dimensional computed tomography angiography (4D-CTA) images for automatic detection of large vessel occlusion (LVO) in the anterior circulation that cause acute ischemic stroke.

METHODS: This retrospective study included 104 LVO patients and 105 non-LVO patients for deep learning models development. Another 30 LVO patients and 31 non-LVO patients formed the time-independent validation set. Four phases of 4D-CTA (arterial phase P1, arterial-venous phase P2, venous phase P3 and late venous phase P4) were arranged and combined and two input methods was used: combined input and superimposed input. Totally 26 models were constructed using a modified HRNet network. Assessment metrics included the areas under the curve (AUC), accuracy, sensitivity, specificity and F1 score. Kappa analysis was performed to assess inter-rater agreement between the best model and radiologists of different seniority.

RESULTS: The P1 + P2 model (combined input) had the best diagnostic performance. In the internal validation set, the AUC was 0.975 (95%CI: 0.878-0.999), accuracy was 0.911, sensitivity was 0.889, specificity was 0.944, and the F1 score was 0.909. In the time-independent validation set, the model demonstrated consistently high performance with an AUC of 0.942 (95%CI: 0.851-0.986), accuracy of 0.902, sensitivity of 0.867, specificity of 0.935, and an F1 score of 0.901. The best model showed strong consistency with the diagnostic efficacy of three radiologists of different seniority (k = 0.84, 0.80, 0.70, respectively).

CONCLUSION: The deep learning model, using combined arterial and arterial-venous phase, was highly effective in detecting LVO, alerting radiologists to speed up the diagnosis.

PMID:38660224 | PMC:PMC11039833 | DOI:10.3389/fnins.2024.1329718

PeerJ Comput Sci. 2024 Mar 27;10:e1937. doi: 10.7717/peerj-cs.1937. eCollection 2024.

ABSTRACT

This article addresses the evolving landscape of data advertising within network-based new media, seeking to mitigate the accuracy limitations prevalent in traditional film and television advertising evaluations. To overcome these challenges, a novel data-driven nonlinear dynamic neural network planning approach is proposed. Its primary objective is to augment the real-time evaluation precision and accuracy of film and television advertising in the dynamic interactive realm of network media. The methodology primarily revolves around formulating a design model for visual advertising in film and television, customized for the dynamic interactive milieu of network media. Leveraging DeepFM+long short-term memory (LSTM) modules in deep learning neural networks, the article embarks on constructing a comprehensive information statistics and data interest model derived from two public datasets. It further engages in feature engineering for visual advertising, crafting self-learning association rules that guide the data-driven design process and system flow. The article concludes by benchmarking the proposed visual neural network model against other models, using F1 and root mean square error (RMSE) metrics for evaluation. The findings affirm that the proposed model, capable of handling dynamic interactions among images, visual text, and more, excels in capturing nonlinear and feature-mining aspects. It exhibits commendable robustness and generalization capabilities within various contexts.

PMID:38660215 | PMC:PMC11041990 | DOI:10.7717/peerj-cs.1937

PeerJ Comput Sci. 2024 Mar 11;10:e1902. doi: 10.7717/peerj-cs.1902. eCollection 2024.

ABSTRACT

Gastrointestinal diseases cause around two million deaths globally. Wireless capsule endoscopy is a recent advancement in medical imaging, but manual diagnosis is challenging due to the large number of images generated. This has led to research into computer-assisted methodologies for diagnosing these images. Endoscopy produces thousands of frames for each patient, making manual examination difficult, laborious, and error-prone. An automated approach is essential to speed up the diagnosis process, reduce costs, and potentially save lives. This study proposes transfer learning-based efficient deep learning methods for detecting gastrointestinal disorders from multiple modalities, aiming to detect gastrointestinal diseases with superior accuracy and reduce the efforts and costs of medical experts. The Kvasir eight-class dataset was used for the experiment, where endoscopic images were preprocessed and enriched with augmentation techniques. An EfficientNet model was optimized via transfer learning and fine tuning, and the model was compared to the most widely used pre-trained deep learning models. The model's efficacy was tested on another independent endoscopic dataset to prove its robustness and reliability.

PMID:38660212 | PMC:PMC11041956 | DOI:10.7717/peerj-cs.1902

PeerJ Comput Sci. 2024 Apr 2;10:e1969. doi: 10.7717/peerj-cs.1969. eCollection 2024.

ABSTRACT

The stock market serves as a macroeconomic indicator, and stock price forecasting aids investors in analysing market trends and industry dynamics. Several deep learning network models have been proposed and extensively applied for stock price prediction and trading scenarios in recent times. Although numerous studies have indicated a significant correlation between market sentiment and stock prices, the majority of stock price predictions rely solely on historical indicator data, with minimal effort to incorporate sentiment analysis into stock price forecasting. Additionally, many deep learning models struggle with handling the long-distance dependencies of large datasets. This can cause them to overlook unexpected stock price fluctuations that may arise from long-term market sentiment, making it challenging to effectively utilise long-term market sentiment information. To address the aforementioned issues, this investigation suggests implementing a new technique called Long-term Sentiment Change Enhanced Temporal Analysis (LEET) which effectively incorporates long-term market sentiment and enhances the precision of stock price forecasts. The LEET method proposes two market sentiment index estimation methods: Exponential Weighted Sentiment Analysis (EWSA) and Weighted Average Sentiment Analysis (WASA). These methods are utilized to extract the market sentiment index. Additionally, the study proposes a Transformer architecture based on ProbAttention with rotational position encoding for enhanced positional information capture of long-term emotions. The LEET methodology underwent validation using the Standard & Poor's 500 (SP500) and FTSE 100 indices. These indices accurately reflect the state of the US and UK equity markets, respectively. The experimental results obtained from a genuine dataset demonstrate that this method is superior to the majority of deep learning network architectures when it comes to predicting stock prices.

PMID:38660208 | PMC:PMC11041952 | DOI:10.7717/peerj-cs.1969

PeerJ Comput Sci. 2024 Mar 27;10:e1925. doi: 10.7717/peerj-cs.1925. eCollection 2024.

ABSTRACT

This article introduces a recognition system for handwritten text in the Pashto language, representing the first attempt to establish a baseline system using the Pashto Handwritten Text Imagebase (PHTI) dataset. Initially, the PHTI dataset underwent pre-processed to eliminate unwanted characters, subsequently, the dataset was divided into training 70%, validation 15%, and test sets 15%. The proposed recognition system is based on multi-dimensional long short-term memory (MD-LSTM) networks. A comprehensive empirical analysis was conducted to determine the optimal parameters for the proposed MD-LSTM architecture; Counter experiments were used to evaluate the performance of the proposed system comparing with the state-of-the-art models on the PHTI dataset. The novelty of our proposed model, compared to other state of the art models, lies in its hidden layer size (i.e., 10, 20, 80) and its Tanh layer size (i.e., 20, 40). The system achieves a Character Error Rate (CER) of 20.77% as a baseline on the test set. The top 20 confusions are reported to check the performance and limitations of the proposed model. The results highlight complications and future perspective of the Pashto language towards the digital transition.

PMID:38660206 | PMC:PMC11041937 | DOI:10.7717/peerj-cs.1925

PeerJ Comput Sci. 2024 Apr 3;10:e1981. doi: 10.7717/peerj-cs.1981. eCollection 2024.

ABSTRACT

BACKGROUND: In today's world, numerous applications integral to various facets of daily life include automatic speech recognition methods. Thus, the development of a successful automatic speech recognition system can significantly augment the convenience of people's daily routines. While many automatic speech recognition systems have been established for widely spoken languages like English, there has been insufficient progress in developing such systems for less common languages such as Turkish. Moreover, due to its agglutinative structure, designing a speech recognition system for Turkish presents greater challenges compared to other language groups. Therefore, our study focused on proposing deep learning models for automatic speech recognition in Turkish, complemented by the integration of a language model.

METHODS: In our study, deep learning models were formulated by incorporating convolutional neural networks, gated recurrent units, long short-term memories, and transformer layers. The Zemberek library was employed to craft the language model to improve system performance. Furthermore, the Bayesian optimization method was applied to fine-tune the hyper-parameters of the deep learning models. To evaluate the model's performance, standard metrics widely used in automatic speech recognition systems, specifically word error rate and character error rate scores, were employed.

RESULTS: Upon reviewing the experimental results, it becomes evident that when optimal hyper-parameters are applied to models developed with various layers, the scores are as follows: Without the use of a language model, the Turkish Microphone Speech Corpus dataset yields scores of 22.2 -word error rate and 14.05-character error rate, while the Turkish Speech Corpus dataset results in scores of 11.5 -word error rate and 4.15 character error rate. Upon incorporating the language model, notable improvements were observed. Specifically, for the Turkish Microphone Speech Corpus dataset, the word error rate score decreased to 9.85, and the character error rate score lowered to 5.35. Similarly, the word error rate score improved to 8.4, and the character error rate score decreased to 2.7 for the Turkish Speech Corpus dataset. These results demonstrate that our model outperforms the studies found in the existing literature.

PMID:38660198 | PMC:PMC11041944 | DOI:10.7717/peerj-cs.1981

Plant Methods. 2024 Apr 24;20(1):56. doi: 10.1186/s13007-024-01191-6.

ABSTRACT

BACKGROUND: Traditional method of wood species identification involves the use of hand lens by wood anatomists, which is a time-consuming method that usually identifies only at the genetic level. Computer vision method can achieve "species" level identification but cannot provide an explanation on what features are used for the identification. Thus, in this study, we used computer vision methods coupled with deep learning to reveal interspecific differences between closely related tree species.

RESULT: A total of 850 images were collected from the cross and tangential sections of 15 wood species. These images were used to construct a deep-learning model to discriminate wood species, and a classification accuracy of 99.3% was obtained. The key features between species in machine identification were targeted by feature visualization methods, mainly the axial parenchyma arrangements and vessel in cross section and the wood ray in tangential section. Moreover, the degree of importance of the vessels of different tree species in the cross-section images was determined by the manual feature labeling method. The results showed that vessels play an important role in the identification of Dalbergia, Pterocarpus, Swartzia, Carapa, and Cedrela, but exhibited limited resolutions on discriminating Swietenia species.

CONCLUSION: The research results provide a computer-assisted tool for identifying endangered tree species in laboratory scenarios, which can be used to combat illegal logging and related trade and contribute to the implementation of CITES convention and the conservation of global biodiversity.

PMID:38659006 | DOI:10.1186/s13007-024-01191-6

BMC Oral Health. 2024 Apr 24;24(1):490. doi: 10.1186/s12903-024-04262-1.

ABSTRACT

BACKGROUND: Deep learning model trained on a large image dataset, can be used to detect and discriminate targets with similar but not identical appearances. The aim of this study is to evaluate the post-training performance of the CNN-based YOLOv5x algorithm in the detection of white spot lesions in post-orthodontic oral photographs using the limited data available and to make a preliminary study for fully automated models that can be clinically integrated in the future.

METHODS: A total of 435 images in JPG format were uploaded into the CranioCatch labeling software and labeled white spot lesions. The labeled images were resized to 640 × 320 while maintaining their aspect ratio before model training. The labeled images were randomly divided into three groups (Training:349 images (1589 labels), Validation:43 images (181 labels), Test:43 images (215 labels)). YOLOv5x algorithm was used to perform deep learning. The segmentation performance of the tested model was visualized and analyzed using ROC analysis and a confusion matrix. True Positive (TP), False Positive (FP), and False Negative (FN) values were determined.

RESULTS: Among the test group images, there were 133 TPs, 36 FPs, and 82 FNs. The model's performance metrics include precision, recall, and F1 score values of detecting white spot lesions were 0.786, 0.618, and 0.692. The AUC value obtained from the ROC analysis was 0.712. The mAP value obtained from the Precision-Recall curve graph was 0.425.

CONCLUSIONS: The model's accuracy and sensitivity in detecting white spot lesions remained lower than expected for practical application, but is a promising and acceptable detection rate compared to previous study. The current study provides a preliminary insight to further improved by increasing the dataset for training, and applying modifications to the deep learning algorithm.

CLINICAL REVELANCE: Deep learning systems can help clinicians to distinguish white spot lesions that may be missed during visual inspection.

PMID:38658959 | DOI:10.1186/s12903-024-04262-1

BMC Genomics. 2024 Apr 24;25(Suppl 1):401. doi: 10.1186/s12864-024-10251-z.

ABSTRACT

BACKGROUND: Most of the important biological mechanisms and functions of transmembrane proteins (TMPs) are realized through their interactions with non-transmembrane proteins(nonTMPs). The interactions between TMPs and nonTMPs in cells play vital roles in intracellular signaling, energy metabolism, investigating membrane-crossing mechanisms, correlations between disease and drugs.

RESULTS: Despite the importance of TMP-nonTMP interactions, the study of them remains in the wet experimental stage, lacking specific and comprehensive studies in the field of bioinformatics. To fill this gap, we performed a comprehensive statistical analysis of known TMP-nonTMP interactions and constructed a deep learning-based predictor to identify potential interactions. The statistical analysis describes known TMP-nonTMP interactions from various perspectives, such as distributions of species and protein families, enrichment of GO and KEGG pathways, as well as hub proteins and subnetwork modules in the PPI network. The predictor implemented by an end-to-end deep learning model can identify potential interactions from protein primary sequence information. The experimental results over the independent validation demonstrated considerable prediction performance with an MCC of 0.541.

CONCLUSIONS: To our knowledge, we were the first to focus on TMP-nonTMP interactions. We comprehensively analyzed them using bioinformatics methods and predicted them via deep learning-based solely on their sequence. This research completes a key link in the protein network, benefits the understanding of protein functions, and helps in pathogenesis studies of diseases and associated drug development.

PMID:38658824 | DOI:10.1186/s12864-024-10251-z

Eye (Lond). 2024 Apr 24. doi: 10.1038/s41433-024-03093-2. Online ahead of print.

ABSTRACT

BACKGROUND/OBJECTIVES: To evaluate the diagnostic performance of B-scan kinetic ultrasonography (USG), standard ultra-widefield (UWF) imaging, and indirect ophthalmoscopy (IDO) in retinal break detection in cataractous eyes.

SUBJECTS/METHODS: We consecutively enrolled 126 cataract patients (including 246 eyes) with no comorbidities that could decrease best corrected visual acuity (BCVA). Three index tests (USG, nonmydriatic UWF, and mydriatic IDO) were performed preoperatively to screen for retinal breaks. One week after cataract extraction, a dilated IDO examination was repeated for the definitive diagnosis of retinal break as the reference standard. The sensitivity, specificity, Youden index (YI), and predictive values of each index test were calculated according to postoperative ophthalmoscopic findings. A deep-learning nomogram was developed to quantify the risk of retinal break presence using patients' baseline data and findings reported from preoperative ophthalmic tests.

RESULTS: Fifty-two eyes (21%) were excluded from appropriate preoperative UWF imaging because of massive lens opacity. The BCVA cutoff point with maximum YI indicating UWF applicability was 0.6 logMAR (YI = 0.3; area under curve [AUC] = 0.7). Among all 246 eyes, preoperative IDO, USG, and UWF showed fair interobserver agreement (all κ > 0.2). According to postoperative IDO findings, the index tests with the highest sensitivity and specificity were USG (100%) and preoperative IDO (99%), respectively.

CONCLUSIONS: For cataractous eyes without vision-impairing comorbidities, a BCVA better than 0.6 logMAR (Snellen acuity, 20/80) allows for appropriate nonmydriatic standard UWF imaging. In a high-volume clinic equipped with skilled ophthalmic examiners, screening with USG followed by directed IDO allows the efficient identification of retinal breaks in cataractous eyes.

PMID:38658680 | DOI:10.1038/s41433-024-03093-2

Sci Rep. 2024 Apr 24;14(1):9465. doi: 10.1038/s41598-024-59622-2.

ABSTRACT

A poor nutritional status is associated with worse pulmonary function and survival in people with cystic fibrosis (pwCF). CF transmembrane conductance regulator modulators can improve pulmonary function and body weight, but more data is needed to evaluate its effects on body composition. In this retrospective study, a pre-trained deep-learning network was used to perform a fully automated body composition analysis on chest CTs from 66 adult pwCF before and after receiving elexacaftor/tezacaftor/ivacaftor (ETI) therapy. Muscle and adipose tissues were quantified and divided by bone volume to obtain body size-adjusted ratios. After receiving ETI therapy, marked increases were observed in all adipose tissue ratios among pwCF, including the total adipose tissue ratio (+ 46.21%, p < 0.001). In contrast, only small, but statistically significant increases of the muscle ratio were measured in the overall study population (+ 1.63%, p = 0.008). Study participants who were initially categorized as underweight experienced more pronounced effects on total adipose tissue ratio (p = 0.002), while gains in muscle ratio were equally distributed across BMI categories (p = 0.832). Our findings suggest that ETI therapy primarily affects adipose tissues, not muscle tissue, in adults with CF. These effects are primarily observed among pwCF who were initially underweight. Our findings may have implications for the future nutritional management of pwCF.

PMID:38658613 | DOI:10.1038/s41598-024-59622-2

Nat Commun. 2024 Apr 24;15(1):3451. doi: 10.1038/s41467-024-47391-5.

ABSTRACT

Enhancers are fast-evolving genomic sequences that control spatiotemporal gene expression patterns. By examining enhancer turnover across mammalian species and in multiple tissue types, we uncover a relationship between the emergence of enhancers and genome organization as a function of germline DNA replication time. While enhancers are most abundant in euchromatic regions, enhancers emerge almost twice as often in late compared to early germline replicating regions, independent of transposable elements. Using a deep learning sequence model, we demonstrate that new enhancers are enriched for mutations that alter transcription factor (TF) binding. Recently evolved enhancers appear to be mostly neutrally evolving and enriched in eQTLs. They also show more tissue specificity than conserved enhancers, and the TFs that bind to these elements, as inferred by binding sequences, also show increased tissue-specific gene expression. We find a similar relationship with DNA replication time in cancer, suggesting that these observations may be time-invariant principles of genome evolution. Our work underscores that genome organization has a profound impact in shaping mammalian gene regulation.

PMID:38658544 | DOI:10.1038/s41467-024-47391-5

Nat Commun. 2024 Apr 24;15(1):3470. doi: 10.1038/s41467-024-47171-1.

ABSTRACT

Identifying active compounds for a target is a time- and resource-intensive task in early drug discovery. Accurate bioactivity prediction using morphological profiles could streamline the process, enabling smaller, more focused compound screens. We investigate the potential of deep learning on unrefined single-concentration activity readouts and Cell Painting data, to predict compound activity across 140 diverse assays. We observe an average ROC-AUC of 0.744 ± 0.108 with 62% of assays achieving ≥0.7, 30% ≥0.8, and 7% ≥0.9. In many cases, the high prediction performance can be achieved using only brightfield images instead of multichannel fluorescence images. A comprehensive analysis shows that Cell Painting-based bioactivity prediction is robust across assay types, technologies, and target classes, with cell-based assays and kinase targets being particularly well-suited for prediction. Experimental validation confirms the enrichment of active compounds. Our findings indicate that models trained on Cell Painting data, combined with a small set of single-concentration data points, can reliably predict the activity of a compound library across diverse targets and assays while maintaining high hit rates and scaffold diversity. This approach has the potential to reduce the size of screening campaigns, saving time and resources, and enabling primary screening with more complex assays.

PMID:38658534 | DOI:10.1038/s41467-024-47171-1

Med Biol Eng Comput. 2024 Apr 25. doi: 10.1007/s11517-024-03092-1. Online ahead of print.

ABSTRACT

The assessment of deformable registration uncertainty is an important task for the safety and reliability of registration methods in clinical applications. However, it is typically done by a manual and time-consuming procedure. We propose a novel automatic method to predict registration uncertainty based on multi-category features and supervised learning. Three types of features, including deformation field statistical features, deformation field physiologically realistic features, and image similarity features, are introduced and calculated to train the random forest regressor for local registration uncertain prediction. Deformation field statistical features represent the numerical stability of registration optimization, which are correlated to the uncertainty of deformation fields; deformation field physiologically realistic features represent the biomechanical properties of organ motions, which mathematically reflect the physiological reality of deformation; image similarity features reflect the similarity between the warped image and fixed image. The multi-category features comprehensively reflect the registration uncertainty. The strategy of spatial adaptive random perturbations is also introduced to accurately simulate spatial distribution of registration uncertainty, which makes deformation field statistical features more discriminative to the uncertainty of deformation fields. Experiments were conducted on three publicly available thoracic CT image datasets. Seventeen randomly selected image pairs are used to train the random forest model, and 9 image pairs are used to evaluate the prediction model. The quantitative experiments on lung CT images show that the proposed method outperforms the baseline method for uncertain prediction of classical iterative optimization-based registration and deep learning-based registration with different registration qualities. The proposed method achieves good performance for registration uncertain prediction, which has great potential in improving the accuracy of registration uncertain prediction.

PMID:38658497 | DOI:10.1007/s11517-024-03092-1

Skeletal Radiol. 2024 Apr 24. doi: 10.1007/s00256-024-04690-8. Online ahead of print.

ABSTRACT

OBJECTIVES: To assess a deep learning-based reconstruction algorithm (DLRecon) in zero echo-time (ZTE) MRI of the shoulder at 1.5 Tesla for improved delineation of osseous findings.

METHODS: In this retrospective study, 63 consecutive exams of 52 patients (28 female) undergoing shoulder MRI at 1.5 Tesla in clinical routine were included. Coronal 3D isotropic radial ZTE pulse sequences were acquired in the standard MR shoulder protocol. In addition to standard-of-care (SOC) image reconstruction, the same raw data was reconstructed with a vendor-supplied prototype DLRecon algorithm. Exams were classified into three subgroups: no pathological findings, degenerative changes, and posttraumatic changes, respectively. Two blinded readers performed bone assessment on a 4-point scale (0-poor, 3-perfect) by qualitatively grading image quality features and delineation of osseous pathologies including diagnostic confidence in the respective subgroups. Quantitatively, signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of bone were measured. Qualitative variables were compared using the Wilcoxon signed-rank test for ordinal data and the McNemar test for dichotomous variables; quantitative measures were compared with Student's t-testing.

RESULTS: DLRecon scored significantly higher than SOC in all visual metrics of image quality (all, p < 0.03), except in the artifact category (p = 0.37). DLRecon also received superior qualitative scores for delineation of osseous pathologies and diagnostic confidence (p ≤ 0.03). Quantitatively, DLRecon achieved superior CNR (95 CI [1.4-3.1]) and SNR (95 CI [15.3-21.5]) of bone than SOC (p < 0.001).

CONCLUSION: DLRecon enhanced image quality in ZTE MRI and improved delineation of osseous pathologies, allowing for increased diagnostic confidence in bone assessment.

PMID:38658419 | DOI:10.1007/s00256-024-04690-8

Lancet Digit Health. 2024 Apr 23:S2589-7500(24)00061-X. doi: 10.1016/S2589-7500(24)00061-X. Online ahead of print.

ABSTRACT

With the rapid growth of interest in and use of large language models (LLMs) across various industries, we are facing some crucial and profound ethical concerns, especially in the medical field. The unique technical architecture and purported emergent abilities of LLMs differentiate them substantially from other artificial intelligence (AI) models and natural language processing techniques used, necessitating a nuanced understanding of LLM ethics. In this Viewpoint, we highlight ethical concerns stemming from the perspectives of users, developers, and regulators, notably focusing on data privacy and rights of use, data provenance, intellectual property contamination, and broad applications and plasticity of LLMs. A comprehensive framework and mitigating strategies will be imperative for the responsible integration of LLMs into medical practice, ensuring alignment with ethical principles and safeguarding against potential societal risks.

PMID:38658283 | DOI:10.1016/S2589-7500(24)00061-X

IEEE Trans Med Imaging. 2024 Apr 24;PP. doi: 10.1109/TMI.2024.3391215. Online ahead of print.

ABSTRACT

Self-supervised learning (SSL) has emerged as a powerful technique for improving the efficiency and effectiveness of deep learning models. Contrastive methods are a prominent family of SSL that extract similar representations of two augmented views of an image while pushing away others in the representation space as negatives. However, the state-of-the-art contrastive methods require large batch sizes and augmentations designed for natural images that are impractical for 3D medical images. To address these limitations, we propose a new longitudinal SSL method, 3DTINC, based on non-contrastive learning. It is designed to learn perturbation-invariant features for 3D optical coherence tomography (OCT) volumes, using augmentations specifically designed for OCT. We introduce a new non-contrastive similarity loss term that learns temporal information implicitly from intra-patient scans acquired at different times. Our experiments show that this temporal information is crucial for predicting progression of retinal diseases, such as age-related macular degeneration (AMD). After pretraining with 3DTINC, we evaluated the learned representations and the prognostic models on two large-scale longitudinal datasets of retinal OCTs where we predict the conversion to wet-AMD within a six-month interval. Our results demonstrate that each component of our contributions is crucial for learning meaningful representations useful in predicting disease progression from longitudinal volumetric scans.

PMID:38656867 | DOI:10.1109/TMI.2024.3391215