Search Results (408)

Pulmonary neuroendocrine proliferations and neoplasms represent a broad spectrum of diseases, ranging from neuroendocrine cell hyperplasia and tumorlets to carcinoid tumors. Carcinoid tumorlets are most commonly located in the peripheral airways and are often incidentally detected as pulmonary micronodules on chest CT. We report the radiological, bronchoscopic, and pathological findings of a case of carcinoid tumorlets presenting as endobronchial nodules in the left main bronchus. The patient had previously undergone a left lower lobectomy five years earlier for a typical carcinoid tumor. Follow-up imaging revealed new endobronchial nodules, which were subsequently confirmed as carcinoid tumorlets through histopathologic analysis. This case highlights the rare presentation of carcinoid tumorlets in the central airways, emphasizing the importance of recognizing their potential for late recurrence and atypical localization. It underscores the necessity for physicians to be aware that pulmonary neuroendocrine tumors can recur over the long term and may present in a multicentric fashion within the disease spectrum. Full article

In this study, we present a comparative analysis of gradient descent-based optimizers frequently used in Convolutional Neural Networks (CNNs), including SGD, mSGD, RMSprop, Adadelta, Nadam, Adamax, Adam, and the recent EVE optimizer. To explore the interaction between optimization strategies and activation functions, we systematically evaluate all combinations of these optimizers with four activation functions—ReLU, LeakyReLU, Tanh, and GELU—across three benchmark image classification datasets: CIFAR-10, Fashion-MNIST (F-MNIST), and Labeled Faces in the Wild (LFW). Each configuration was assessed using multiple evaluation metrics, including accuracy, precision, recall, F1-score, mean absolute error (MAE), and mean squared error (MSE). All experiments were performed using k-fold cross-validation to ensure statistical robustness. Additionally, two-way ANOVA was employed to validate the significance of differences across optimizer–activation combinations. This study aims to highlight the importance of jointly selecting optimizers and activation functions to enhance training dynamics and generalization in CNNs. We also consider the role of critical hyperparameters, such as learning rate and regularization methods, in influencing optimization stability. This work provides valuable insights into the optimizer–activation interplay and offers practical guidance for improving architectural and hyperparameter configurations in CNN-based deep learning models. Full article

The prediction of anthropometric measurements from 2D body images, particularly for children, remains an under-explored area despite its potential applications in healthcare, fashion, and fitness. While pose estimation and body shape classification have garnered extensive attention, estimating body measurements and body mass index (BMI) from images presents unique challenges and opportunities. This paper provides a comprehensive review of the current methodologies, focusing on deep-learning approaches, both standalone and in combination with traditional machine-learning techniques, for inferring body measurements from facial and full-body images. We discuss the strengths and limitations of commonly used datasets, proposing the need for more inclusive and diverse collections to improve model performance. Our findings indicate that deep-learning models, especially when combined with traditional machine-learning techniques, offer the most accurate predictions. We further highlight the promise of vision transformers in advancing the field while stressing the importance of addressing model explainability. Finally, we evaluate the current state of the field, comparing recent results and focusing on the deviations from ground truth, ultimately providing recommendations for future research directions. Full article

Background: While magnetic resonance imaging (MRI) is commonly used to identify meniscal tears, intraoperative assessment typically dictates repairability. This study evaluated whether a simplified MRI-based scoring system could reliably predict meniscal repair versus meniscectomy. Methods: Patients who underwent meniscectomy or meniscal repair between 2010 and 2018 were retrospectively identified. Preoperative MRIs were independently reviewed in a blinded fashion by two radiologists and one orthopedic sports surgeon. Reviewers scored images based on four arthroscopic criteria for tear repairability, with one point awarded for each of the following criteria—(1) proximity within 4 mm of the meniscosynovial junction, (2) length > 10 mm, (3) presence of intact inner meniscal segment, and (4) >50% meniscal thickness. Tears scoring four points were considered repairable. Accuracy, sensitivity, and positive and negative predictive values were calculated against the actual procedure performed. Inter- and intraobserver reliability were evaluated using kappa statistics. The predictive performance of each individual criterion was also analyzed. Results: A total of 202 meniscal tears were included (134 meniscectomies and 68 repairs). Reviewer accuracy in predicting repairability ranged from 48% to 76%. Intraobserver reliability was moderate to substantial (κ = 0.42–0.66), whereas interobserver reliability was poor to moderate (pairwise κ = 0.07–0.43; Fleiss’ κ = 0.11). Analysis of individual MRI criteria demonstrated limited predictive value, with most criteria achieving less than 50% accuracy across reviewers. Conclusions: MRI-based prediction of meniscal repairability using arthroscopic criteria demonstrated limited accuracy and poor interobserver reliability. Overall predictive reliability remains insufficient for clinical decision-making. Further investigation, integrating advanced imaging techniques and artificial intelligence, may improve the preoperative assessment of meniscal repairability. Full article

Book covers are often the first component seen before a reader engages with a book’s contents; therefore, careful consideration is given to the text and image features that constitute their design. This study investigates the effects of the presentation of verbal (text) and visual (image) features on memorability and aesthetic evaluation in the context of book covers. To this aim, 50 participants took part in a memory recognition task in which the same book cover information was encoded in a learning phase, and either text or image features from the book covers acted as an informational cue for memory recognition and aesthetic evaluations. Our results revealed that image features significantly aided memory performance more than text features. Image features that were rated more beautiful were not better recognized as a result. However, differences in memory performance were found in relation to familiarity and, in a non-linear fashion, the extent to which the book’s contents could be inferred from the image’s informational content. Additionally, reading behavior was not found to influence memory performance. These results are discussed with regard to the interplay of text and image informational cues on book cover perception and provide implications for future studies. Full article

Pulse-coupled neural networks (PCNNs) are capable of segmenting digital images in a multistage unsupervised fashion; however, optimal output selection remains challenging. To address the above problem, this paper emphasizes the role of the step size, which influences the decreasing speed of the membrane potential and the dynamic threshold profoundly. A dynamic-step-size mechanism is proposed, utilizing trigonometric functions to adaptively control segmentation granularity, along with the supervised optimization of a single parameter $\varphi$ via intersection over union (IoU) maximization, reducing tuning complexity. Thus, the number of groups of image segmentation becomes controllable and the model itself becomes more adaptive than ever for various scenarios. Experimental results further demonstrate the enhanced robustness under noise (92.1% Dice at $\sigma =0.2$), outperforming SPCNN and PCNN with IoU = 0.8863, Dice = 0.901, and 0.8684 s/image. Full article

Background/Objectives: Convolutional Neural Networks (CNNs), while effective in tasks such as image classification and language processing, often experience overfitting and inefficient training due to static, structure-agnostic regularization techniques like traditional dropout. This study aims to address these limitations by proposing a more dynamic and context-sensitive dropout strategy. Methods: We introduce Probabilistic Feature Importance Dropout (PFID), a novel regularization method that assigns dropout rates based on the probabilistic significance of individual features. PFID is integrated with adaptive, structured, and contextual dropout strategies, forming a unified framework for intelligent regularization. Results: Experimental evaluation on standard benchmark datasets including CIFAR-10, MNIST, and Fashion MNIST demonstrated that PFID significantly improves performance metrics such as classification accuracy, training loss, and computational efficiency compared to conventional dropout methods. Conclusions: PFID offers a practical and scalable solution for enhancing CNN generalization and training efficiency. Its dynamic nature and feature-aware design provide a strong foundation for future advancements in adaptive regularization for deep learning models. Full article

Since the 1960s, Iran’s major cities have experienced significant migration from the country’s rural areas and from other nations. Although many urban planning and design concepts can be traced back to Iran, the country’s planning machinery has failed to effectively regulate urban growth, notably in the city peripheries, where land use has changed radically as semi-rural areas have been developed in a haphazard fashion with scant adherence to existing plans and planning regulations. Farahzad is one such area in the urban periphery of Tehran, where a range of sub-standard dwellings have been built, and urban services are deficient in many regards. This article examines how the urban landscape has evolved, how the resident population has grown, and the nature of the social and economic issues that persist today. The research method combines an analysis of the extant literature and local authority documentation, images developed from GIS data, and first-hand interviews with local practitioners to explore the growth of the neighbourhood in recent decades and assess the current problems confronting both residents and local authorities. The novelty of this article lies in the use of GIS-generated images and urban fabric classifications to assess the growth of the neighbourhood since the turn of century, during which time the planning machinery has generally failed to provide an adequate framework for development in this area of the Tehran urban periphery. Indeed, findings suggest that land use zoning has played little part in guiding or controlling urban development in Farahzad, and that identifying urban fabrics may prove a useful way of assessing socio-economic and physical development needs in such circumstances. This article makes a small contribution to our understanding of the change dynamics in a peripheral neighbourhood of a major city in the developing world. Full article

We developed a shoe recommendation system that integrates generative artificial intelligence (AI) and convolutional neural networks (CNNs) to enhance image recognition and personalize recommendations. The system utilizes CNNs to accurately identify shoe types from user-uploaded images. Utilizing the capabilities of generative AI, the system generates custom shoe suggestions based on weather and location. The proposed system minimizes the need for manual searching but enhances user experience by providing an efficient, automated, and visually driven solution for selecting shoes. The effectiveness of integrating image recognition and generative techniques paves the way for advancements in AI-driven fashion recommendation systems. The developed method offers a powerful tool for increasing customer engagement and satisfaction by delivering personalized and fashion-forward shoe recommendations. Full article

Inkjet-printed circuits on flexible substrates are rapidly emerging as a key technology in flexible electronics, driven by their minimal fabrication process, cost-effectiveness, and environmental sustainability. Recent advancements in inkjet-printed devices and circuits have broadened their applications in both sensing and computing. Building on this progress, this work has developed a nonlinear computational element coined as mTanh to serve as an activation function in neural networks. Activation functions are essential in neural networks as they introduce nonlinearity, enabling machine learning models to capture complex patterns. However, widely used functions such as Tanh and sigmoid often suffer from the vanishing gradient problem, limiting the depth of neural networks. To address this, alternative functions like ReLU and Leaky ReLU have been explored, yet these also introduce challenges such as the dying ReLU issue, bias shifting, and noise sensitivity. The proposed mTanh activation function effectively mitigates the vanishing gradient problem, allowing for the development of deeper neural network architectures without compromising training efficiency. This study demonstrates the feasibility of mTanh as an activation function by integrating it into an Echo State Network to predict the Mackey–Glass time series signal. The results show that mTanh performs comparably to Tanh, ReLU, and Leaky ReLU in this task. Additionally, the vanishing gradient resistance of the mTanh function was evaluated by implementing it in a deep multi-layer perceptron model for Fashion MNIST image classification. The study indicates that mTanh enables the addition of 3–5 extra layers compared to Tanh and sigmoid, while exhibiting vanishing gradient resistance similar to ReLU. These results highlight the potential of mTanh as a promising activation function for deep learning models, particularly in flexible electronics applications. Full article

Pose transfer methods often struggle to simultaneously preserve fine-grained clothing textures and facial details, especially under large pose variations. To address these limitations, we propose a model based on the Multi-scale attention guided pose transfer model, with modifications to its residual block by integrating the convolutional block attention module and changing the activation function from ReLU to Mish to capture more features related to clothing and skin color. Additionally, as the generated images had facial features differing from the original image, we propose two different facial feature loss functions to help the model learn more precise image features. According to the experimental results, the proposed method demonstrates superior performance compared to the Multi-scale Attention Guided Pose Transfer (MAGPT) on the DeepFashion dataset, achieving a 3.41% reduction in FID, a 0.65% improvement in SSIM, a 2% decrease in LPIPS, and a 2.7% decrease in LPIPS. Ultimately, only one reference image is required to enable users to transform into different action videos with the proposed system architecture. Full article

Microwave computational imaging (MCI) based on coded apertures does not rely on relative motion between the radar platform and the target, enabling forward-looking imaging. The performance of MCI depends on the computational methods and modulation of the coded aperture, particularly its design. However, current research methods treat the optimization of the coded aperture and computational imaging processing as independent tasks, with no unified framework to link these two aspects, limiting the potential for improving system performance. This paper proposes a novel deep learning-based MCI framework that jointly optimizes the coded aperture and image reconstruction process. Unlike traditional methods that decouple these two stages, our approach trains the sensing and reconstruction networks in an end-to-end fashion. The key novelty lies in constructing an end-to-end imaging network based on a convolutional neural network (CNN) where the coded aperture is modeled as a convolutional layer within the network. Physical constraints on the coded aperture are enforced by adding regularizers to the loss function. Simulation experiments demonstrate that under low signal-to-noise ratio (SNR) and low compression ratio conditions, the proposed method improves peak signal-to-noise ratio (PSNR) by 5 dB to 8 dB, enhances SSIM by 10% to 15%, and reduces relative imaging error by 0.5% to 1%. Full article

Introduction: Neurodegenerative diseases like progressive supranuclear palsy (PSP) present challenges concerning their diagnosis. Neuroimaging using magnetic resonance (MRI) may add diagnostic value. However, modern techniques such as volumetric assessment using Voxel-Based Morphometry (VBM), although proven to be more accurate and superior compared to MRI, have not gained popularity among scientists in the investigation of neurological disorders due to their higher cost and time-consuming applications. Conventional brain MRI methods may present a quick, practical, and easy-to-use imaging rating tool for the differential diagnosis of PSP. The purpose of this study is to evaluate a string of existing visual MRI rating scales and signs regarding their impact for the diagnosis of PSP. Materials and Methods: The population study consisted of 30 patients suffering from PSP and 72 healthy controls. Each study participant underwent a brain MRI, which was subsequently examined by two independent researchers in a double-blinded fashion. Fifteen visual rating scales and signs were evaluated, including pontine atrophy, cerebellar atrophy, midbrain atrophy, aqueduct of Sylvius enlargement, cerebellar peduncle hyperintensities, enlargement of the fourth ventricle (100% sensitivity and 71% specificity) and left temporal lobe atrophy (97% sensitivity and 78% specificity). Conclusions: Enlargement of the Sylvius aqueduct, enlargement of the fourth ventricle and atrophy of both temporal lobes together with the presence of morning glory and hummingbird signs can be easily and quickly distinguished and identified by an experienced radiologist without involving any complex analysis, making them useful tools for PSP diagnosis. MRI visual scale measurements could be added to the diagnostic criteria of PSP and may serve as an alternative to highly technical and more sophisticated quantification methods. Full article

The rapid growth in the global apparel market and the rise of online consumption underscore the necessity for intelligent clothing recommendation systems that balance visual compatibility with personalized preferences, particularly price sensitivity. Existing recommendation systems often neglect nuanced consumer price behaviors, limiting their ability to deliver truly personalized suggestions. To address this gap, we propose DeepFMP, a multimodal deep learning framework that integrates visual, textual, and price features through an enhanced DeepFM architecture. Leveraging the IQON3000 dataset, our model employs ResNet-50 and BERT for image and text feature extraction, alongside a comprehensive price feature module capturing individual, statistical, and category-specific price patterns. An attention mechanism optimizes multimodal fusion, enabling robust modeling of user preferences. Comparative experiments demonstrate that DeepFMP outperforms state-of-the-art baselines (LR, FM, Wide & Deep, GP-BPR, and DeepFM), achieving AUC improvements of 1.6–12.2% and NDCG@10 gains of up to 3.2%. Case analyses further reveal that DeepFMP effectively improves the recommendation accuracy, offering actionable insights for personalized marketing. This work advances multimodal recommendation systems by emphasizing price sensitivity as a pivotal factor, providing a scalable solution for enhancing user satisfaction and commercial efficacy in fashion e-commerce. Full article

Background/Objectives: The morphological characteristics of breast tumors play a crucial role in the preliminary diagnosis of breast cancer. However, malignant tumors often exhibit rough, irregular edges and unclear, boundaries in ultrasound images. Additionally, variations in tumor size, location, and shape further complicate the accurate segmentation of breast tumors from ultrasound images. Methods: For these difficulties, this paper introduces a breast ultrasound tumor segmentation network comprising a multi-scale feature acquisition (MFA) module and a multi-input edge supplement (MES) module. The MFA module effectively incorporates dilated convolutions of various sizes in a serial-parallel fashion to capture tumor features at diverse scales. Then, the MES module is employed to enhance the output of each decoder layer by supplementing edge information. This process aims to improve the overall integrity of tumor boundaries, contributing to more refined segmentation results. Results: The mean Dice (mDice), Pixel Accuracy (PA), Intersection over Union (IoU), Recall, and Hausdorff Distance (HD) of this method for the publicly available breast ultrasound image (BUSI) dataset were 79.43%, 96.84%, 83.00%, 87.17%, and 19.71 mm, respectively, and for the dataset of Fujian Cancer Hospital, 90.45%, 97.55%, 90.08%, 93.72%, and 11.02 mm, respectively. In the BUSI dataset, compared to the original UNet, the Dice for malignant tumors increased by 14.59%, and the HD decreased by 17.13 mm. Conclusions: Our method is capable of accurately segmenting breast tumor ultrasound images, which provides very valuable edge information for subsequent diagnosis of breast cancer. The experimental results show that our method has made substantial progress in improving accuracy. Full article