Introduction Myocardial perfusion imaging (MPI) using single-photon emission computed tomography (SPECT) is widely used for diagnosing and risk-stratifying coronary artery disease (CAD). Accurate myocardial segmentation is essential for reliable quantification of left ventricular function; however, manual delineation remains time-consuming and prone to inter-observer variability. This study systematically compares five U-Net encoder backbones (CNN, MobileNet, VGG-Attention, Inception, and ResNet50-Attention) and proposes a novel training-free weighted ensemble strategy suitable for real-time clinical deployment. Methods Two MPI SPECT datasets were analyzed (small: 609 images; large: 5,169 images), each paired with expert-verified ground truth masks. Images were preprocessed to 64×64 pixels and augmented using flipping, grid distortion, Gaussian noise injection, and contrast adjustments. All models were trained under identical hyperparameter settings and evaluated using mIoU, Dice coefficient, pixel accuracy, and AUC. K-fold cross-validation (K=5) and an ablation study comparing cross-entropy and Tversky loss functions were performed to assess statistical robustness and segmentation balance. Results The Inception-based U-Net achieved the highest single-model performance on both datasets (small: Dice 0.9656, mIoU 0.9358; large: Dice 0.9538, mIoU 0.9197). The proposed weighted ensemble consistently outperformed all individual architectures (small: Dice 0.9679, mIoU 0.9398, AUC 0.9994, pixel accuracy 99.45%; large: Dice 0.9544, mIoU 0.9205, AUC 0.9979, pixel accuracy 99.25%), achieving near-zero false-positive rates and the lowest false-negative rates across both datasets. K-fold cross-validation confirmed improved foreground detection robustness over holdout splitting. MobileNet offers the best computational efficiency (training time: 3.71 min; model size: 24.04 MB), representing a viable option for resource-constrained settings. Conclusion The proposed training-free weighted ensemble of U-Net architecture achieves state-of-the-art segmentation accuracy for myocardial perfusion SPECT imaging, with near-perfect AUC and pixel accuracy across datasets of varying scale. These results demonstrate strong potential for clinical integration into nuclear cardiology workflows, reducing dependence on manual segmentation and supporting efficient, reproducible left ventricular quantification.

Condition assessment of stone structures is crucial to maintain their durability. To improve the identification of stone cracks, a triple-stage framework for crack detection, segmentation, and decision-support clustering is proposed. The framework starts with an ensemble of state-of-the-art YOLO models to improve crack detection. The detected crack regions are then fed to an enhanced MobileNetV2U-Net for better crack localization. Thereafter, features are extracted from the detected and segmented stone crack regions, and the K-means and Spectral clustering are utilized to categorize crack patterns. Intensive experiments and detailed comparisons are performed to test the proposed approach. Finally, a user-friendly GUI is designed to simplify the complexity of the proposed framework. Results prove that the YOLO ensemble detector and MobileNetV2U-Net model exhibit the best performances based on statistical metrics. Moreover, it is proven that spectral clustering using five clusters applied to the detected-segmented crack patterns is the best-employed scenario.

Background Patients with breast cancer undergoing biological therapy and/or chemotherapy perform multiple radionuclide angiography (RNA) or multigated acquisition (MUGA) scans to assess cardiotoxicity. The association between RNA imaging parameters and left ventricular (LV) ejection fraction (LVEF) remains unclear. Objectives This study aimed to extract and evaluate the association of several novel imaging biomarkers to detect changes in LVEF in patients with breast cancer undergoing chemotherapy. Methods We developed and optimized a novel set of MATLAB routines called the “RNA Toolbox” to extract parameters from RNA images. The code was optimized using various statistical tests (e.g., ANOVA, Bland-Altman, and intraclass correlation tests). We quantitatively analyzed the images to determine the association between these parameters using regression models and receiver operating characteristic (ROC) curves. Results The code was reproducible and showed good agreement with validated clinical software for the parameters extracted from both packages. The regression model and ROC results were statistically significant in predicting LVEF (R2 = 0.40, P < 0.001) (AUC = 0.78). Some time-based, shape-based, and count-based parameters were significantly associated with post-chemotherapy LVEF (β = 0.09, P < 0.001), LVEF of phase image (β = 4, P = 0.030), approximate entropy (ApEn) (β = 11.6, P = 0.001), ApEn (diastolic and systolic) (β = 39, P = 0.002) and LV systole size (β = 0.03, P = 0.010). Conclusions Despite the limited sample size, we observed evidence of associations between several parameters and LVEF. We believe that these parameters will be more beneficial than the current methods for patients undergoing cardiotoxic chemotherapy. Moreover, this approach can aid physicians in evaluating subclinical cardiac changes during chemotherapy, and in understanding the potential benefits of cardioprotective drugs.

In an era of massive construction, damaged and aging infrastructure are becoming more common. Defects, such as cracking, spalling, etc., are main types of structural damage that widely occur. Hence, ensuring the safe operation of existing infrastructure through health monitoring has emerged as an important challenge facing engineers. In recent years, intelligent approaches, such as data-driven machines and deep learning crack detection have gradually dominated over traditional methods. Among them, the semantic segmentation using deep learning models is a process of the characterization of accurate locations and portraits of cracks using pixel-level classification. Most available studies rely on single-model knowledge to perform this task. However, it is well-known that the single model might suffer from low variance and low ability to generalize in case of data alteration. By leveraging the ensemble deep learning philosophy, a novel collaborative semantic segmentation of concrete cracks method called Co-CrackSegment is proposed. Firstly, five models, namely the U-net, SegNet, DeepCrack19, DeepLabV3-ResNet50, and DeepLabV3-ResNet101 are trained to serve as core models for the ensemble model Co-CrackSegment. To build the ensemble model Co-CrackSegment, a new iterative approach based on the best evaluation metrics, namely the Dice score, IoU, pixel accuracy, precision, and recall metrics is developed. Results show that the Co-CrackSegment exhibits a prominent performance compared with core models and weighted average ensemble by means of the considered best statistical metrics.

Concrete structures when exposed to elevated temperature significantly decline their original properties. High temperatures substantially affect the concrete physical and chemical properties causing significant structural decay and generalized damage impairing the safety and serviceability of the structure. Due to the great importance of concrete behavior at elevated temperatures and under fire, many studies have been conducted on cementitious composites, and the most relevant properties have been studied and evaluated. In particular, fiber-reinforced concrete (FRC) has been a subject of great interest in the last decade due to its superior properties compared to ordinary concrete. Several experimental studies and analytical models have been presented to predict concrete and FRC properties. Among the predictive models, machine learning (ML) tools have shown great merits over other analytical models due to their relative accuracy, generalization abilities, flexible mathematical framework, and cost-effective features. Among the ML, the deep learning (DL) models show remarkable performance when predicting the concrete and the FRC properties at high temperatures because of their ability to deal with more complex nonlinear correlations or difficult regression problems. This review paper presents a pioneering survey of the various ML and DL model implementations predicting concrete and FRC properties at high temperatures. The manuscript aims to establish a solid platform on the state of the art for machine and deep learning prediction of cementitious composites’ properties at elevated temperatures. It aims to provide interested researchers with research indications, directions, challenges, recommendations, and future perspectives.

Background: Myocardial perfusion imaging (MPI) is a type of single-photon emission computed tomography (SPECT) used to evaluate patients with suspected or confirmed coronary artery disease (CAD). Detection and diagnosis of CAD are complex processes requiring precise and accurate image processing. Proper segmentation is critical for accurate diagnosis, but segmentation issues can pose significant challenges, leading to diagnostic difficulties. Machine learning (ML) algorithms have demonstrated superior performance in addressing segmentation problems. Methods: In this study, a deep learning (DL) algorithm, U-Net, was employed to enhance segmentation accuracy for image segmentation in MPI. Data were collected from 1100 patients who underwent MPI studies at Al-Jahra Hospital between 2015 and 2024. To train the U-Net model, 100 studies were segmented by nuclear medicine (NM) experts to create a ground truth (gold-standard coordinates). The dataset was divided into a training set (n = 100 images) and a validation set (n = 900 images). The performance of the U-Net model was evaluated using multiple cross-validation metrics, including accuracy, precision, intersection over union (IOU), recall, and F1 score. Result: A dataset of 4560 images and corresponding masks was generated. Both holdout and k-fold (k = 5) validation strategies were applied, utilizing cross-entropy and Dice score as evaluation metrics. The best results were achieved with the holdout split and cross-entropy loss function, yielding a test accuracy of 98.9%, a test IOU of 89.6%, and a test Dice coefficient of 94%. The k-fold validation scenario provided a more balanced true positive and false positive rate. The U-Net segmentation results were comparable to those produced by expert nuclear medicine technologists, with no significant difference (p = 0.1). Conclusions: The findings demonstrate that the U-Net model effectively addresses some segmentation challenges in MPI, facilitating improved diagnosis and analysis of mega data.

Automatic image-based crack detection of concrete bridge decks contributes to safer bridge operation and bridge health monitoring. Existing models suffer from overfitting and low generalization abilities. Moreover, their performances highly depend on the model architecture, training method, data source, etc. To address these challenges, several hybrid self-designed and transfer learning ensemble models have been introduced for the efficient and accurate intelligent crack detection. Firstly, some self-designed convolutional neural networks (CNNs) are constructed from scratch using labeled crack and non-crack images from modified existing bridge deck image dataset. Secondly, some pretrained transfer learning models, namely the VGG16, VGG19, ResNet50, MobileNetV3Small Model, InceptionResNetV2, EfficientNetV2B0, Xception, and InceptionV3 are adopted to check the efficiency of transfer learning in detecting cracks in bridge decks images. Using the developed CNNs and transfer learning models, several ensemble learning models between the self-designed CNNs, transfer learning CNNs, as well as hybrid self-designed CNNs and transfer learning models are developed. The ensemble learning strategies including the weighted average, stacking, Adaboost, Gradient boosting, and XGBoost ensembles are utilized to construct the ensemble learning models aiming to increase the prediction accuracy and improve generalization ability. Results indicate that the hybrid ensemble learning between the self-designed CNNs and the transfer learning models highly improve the precision and accuracy of the individual models and can be well implemented for image-based bridge deck crack detection.

Early identification of concrete cracks and multi-class detection can help to avoid future deformation or collapse in concrete structures. Available traditional detection and methodologies require enormous effort and time. To overcome such difficulties, current vision-based deep learning models can effectively detect and classify various concrete cracks. This study introduces a novel multi-stage deep learning framework for crack detection and type classification. First, the recently developed YOLOV10 model is trained to detect possible defective regions in concrete images. After that, a modified vision transformer (ViT) model is trained to classify concrete images into three main types: normal, simple cracks, and multi-branched cracks. The evaluation process includes feeding concrete test images into the trained YOLOV10 model, identifying the possible defect regions, and finally delivering the detected regions into the trained ViT model, which decides the appropriate crack type of those detected regions. Experiments are conducted using the individual ViT model and the proposed multi-stage framework. To improve the generation ability, multi-source datasets of concrete structures are used. For the classification part, a concrete crack dataset consisting of 12,000 images of three classes is utilized, while for the detection part, a dataset composed of various materials from historical buildings containing 1116 concrete images with their corresponding bounding boxes, is utilized. Results prove that the proposed multi-stage model accurately classifies crack types with 90.67% precision, 90.03% recall, and 90.34% F1-score. The results also show that the proposed model outperforms the individual classification model by 10.9%, 19.99%, and 19.2% for precision, recall, and F1-score, respectively. The proposed multi-stage YOLOV10-ViT model can be integrated into the construction systems which are based on crack materials to obtain early warning of possible future deformation in concrete structures.

Breast cancer prediction is essential for preventing and treating cancer. In this research, a novel breast cancer prediction model is introduced. In addition, this research aims to provide a range-based cancer score instead of binary classification results (yes or no). The Breast Cancer Surveillance Consortium dataset (BCSC) dataset is used and modified by applying a proposed probabilistic model to achieve the range-based cancer score. The suggested model analyses a sub dataset of the whole BCSC dataset, including 67632 records and 13 risk factors. Three types of statistics are acquired (general cancer and non-cancer probabilities, previous medical knowledge, and the likelihood of each risk factor given all prediction classes). The model also uses the weighting methodology to achieve the best fusion of the BCSC's risk factors. The computation of the final prediction score is done using the post probability of the weighted combination of risk factors and the three statistics acquired from the probabilistic model. This final prediction is added to the BCSC dataset, and the new version of the BCSC dataset is used to train an ensemble model consisting of 30 learners. The experiments are applied using the sub and the whole datasets (including 317880 medical records). The results indicate that the new range-based model is accurate and robust with an accuracy of 91.33%, a false rejection rate of 1.12%, and an AUC of 0.9795. The new version of the BCSC dataset can be used for further research and analysis.

OBJECTIVE:Early prediction of breast cancer is one of the most essential fields of medicine. Many studies have introduced prediction approaches to facilitate the early prediction and estimate the future occurrence based on mammography periodic tests. In the current research, we introduce a novel machine learning tool for the early prediction of breast cancer.METHODS:Three basic resources are used to identify the most essential risk factors; including the BCSC (Breast Cancer Surveillance Consortium) dataset, a medical questionnaire, and multiple international breast cancer reports. The BCSC dataset has been normalized and balanced; consequently, the questionnaire and the medical reports are analyzed in order to define the degree of importance and a potential weight factor of each risk factor. These weights are used to scale risk factors and then the optimizable tree-based ML model is trained using the balanced weighted risk factors datasets.RESULTS:Three balanced versions of the BCSC dataset are used; oversampled, down-sampled and mixed datasets. Each risk factor has a weight (1, 2 or 4) assigned based on a mathematical modelling of the questionnaire and the international breast cancer reports. The experiments are applied on the weighted and non-weighted versions of the database, and they indicate that the performance increases significantly by using the weighted version of the risk factors. The tests prove that the down-weighting of the non-essential risk factor increases the accuracy and reduces errors. The overall accuracy of the weighted balanced datasets reaches 100%, 95.8% and 95.9% for down-sampled, oversampled and mixed datasets respectively.CONCLUSION:Weighting the risk factors of the BCSC dataset improves the performance by increasing the accuracy and reducing the false rejection and false discovery rates for all versions of balanced datasets. The weighting approach can also be used to improve the estimation score of breast cancer by scaling the individual scores of risk factors.

Goal: Medical support systems became very important for the decision-making especially in the current world medical war (Corona Virus). Methods: Lung CT images and textual information of cases collected by the Italian society of medical and interventional radiology site are used to create a textual-image fusion database consisting of 120 cases of COVID19 and other diseases. Machine Learning (ML) algorithm called Random Forests (RF) is applied to design the textual model, while the GoogleNet DL network is used to build the image model. A new fusion technique is used to get the final decision. Results: The medical support system achieves 96.7% accuracy for COVID-19 diagnosis. Conclusions: The designed medical support model can be used for the detection of COVID-19 in order to support the physician's decision especially for people who cannot pay or wait for the Polymerase Chain Reaction (PCR) test results. The model has accurate results with a maximum error of 4.5%.

Background Lung cancer is one of the deadliest cancer in the world. Hundreds of researches are presented annually in the field of lung cancer treatment, diagnosis and early prediction. The current research focuses on the early prediction of lung cancer via analysis of the most dangerous risk factors. Methods A novel tool for the early prediction of lung cancer is designed following three stages: the analysis of an international cancer database, the classification study of the results of local medical questionnaires and the international medical opinion obtained from recently published medical reports. Results The tool is tested using local medical cases and the local medical opinion(s) is (are) used to determine the accuracy of the scores obtained. The Machine Learning approaches are also used to analyze 1000 patient records from an international dataset to compare our results with the international ones. Conclusions The designed tool facilitates computing the risk factors for people who are unable to perform costly hospital tests. It does not require entering all risk inputs and produces the risk factor of lung cancer as a percentage in less than a second. The comparative study with medical opinion and the performance evaluation have confirmed the accuracy of the results.

The current research introduces a novel method for fracture detection and classification. The basic stages of fracture detection includes preprocessing of bone image and morphological operations to obtain the ROI region which is manipulated by a post processing stage to remove non-fracture pixels. The suggested approach extract three features from bone image which are transverse, cracks and divergence features in order to define the fracture type or integrity of bone image. The designed systems detect the different fracture types correctly beside the hybrid fracture type. We applied experiments on a dataset consisting of 155 bone images including 100 fracture images, 30 hand-fingers images and 25 normal images. The systems achieved 92% true detection rate for general bone fractures, 93.33% true detection rate for finger bone fractures and 93.33% true rejection rate. The experiments showed that the inner false detection rates between each type and the others are less than 4%.

ISSN: 2347-8578 www.ijcstjournal.org Page 78 Hybrid Recognition System under Feature Selection and Fusion Eng. Ali M Mayya , Dr. Mariam M Saii [2] PhD Student , Professor Assistance [2] Department of Computer Engineering University of Tishreen, Lattakia Syria ABSTRACT The study suggests a hybrid human recognition system based on face ear and palm print images. The aim of our study is to show the importance of biometric fusion for enhancement of the recognition rate. The system takes face images and segment them into face and ear. Then face, ear and palm images are extracted and fused. The FFBPNN followed by mahalanobis classifier is used in classification stage. The Experiments is applied on 5 databases with different illumination and pose variations and the best result obtained from the face-ear-palm fusion features with 97.5% recognition rate compared with 94.33% for face-ear fusion.

Multimodal biometric systems are more promised and accurate than unimodal ones. Beside increase the performance, the multimodal system minimize the universality problem. The current research introduces a new iris and palm fusion system. The feature of palm print is extracted using connectivity points and lifelines orientations, while features of iris is extracted using wavelet transform. The classification method was the distance classifier. The score level fusion is applied using modified version of majority voter. The system accuracy was 97.29% for palm, 71.97% and 98.54% for fusion.

A new approach of human recognition using ear images is introduced. It consists of two basic steps which are the ear segmentation and ear recognition. In the first one, Likelihood skin detector is used to determine the skin areas in the side face images. Then, some of the morphological operations are applied to determine the ear region. This region is extracted using image processing techniques. The ear recognition step depends on the segmented ear images as inputs. A hybrid PCA_Wavelet algorithm is used to extract the ear features from ear. Finally, the feed forwarding back propagation neural network is trained using the feature vectors. Tests which applied on 460 images, which have been taken during 4 months and under different illumination and pose variations, show that the system achieved a rate of 96.73% for ear extraction and 98.9% for recognition. More experiments are done to specify the best wavelet level, the best number of features, the best classification method, and the best threshold value. The study is also compared with other ones at the area of ear recognition. Correspondence to: Ali Mahmoud Mayya, Computer and Automatic Control Engineering department, Tishreen University, Syria, E-mail: alimia1988@yahoo.com

This paper proposes a new speed approach for the segmentation of the lung images in order to detect and extract the tumor region. The approach consists of two main stages, which are the preprocessing stage, marker watershed stage and the tumor detection stage. The preprocessing consists of laplacian filtering to enhance edges and make the next stages more efficient. The marker watershed step applies the Sobel gradient function on the foreground and background markers to get the possible tumor region. The post processing stage consists of tumor detection and segmentation in which the area of the tumor is calculated. The results are done on a medical lung database obtained from Tishreen hospital (in Lattakia, Syria) which consists of 59 images from 10 persons. The result shows robustness of the system in detecting and segmenting tumor region in different depths. The designed GUI supplies user with tumor region and area, and time of each stage.