Background: Breast cancer, the most common cancer among women, necessitates early detection. Despite advances in Computer-Aided Diagnosis (CAD), lesion detection in mammograms remains challenging. Artificial Intelligence (AI) in radiology offers significant potential to enhance diagnostic accuracy in medical imaging.Objective: This study compares object detection methods to identify the most effective model for smart diagnostic systems. This comprehensive study is the first to apply the advanced You Only Look Once version 12 (YOLO-v12) architecture for the automated detection and localization of lesions in mammographic images and to identify their malignancy or benignity status with high precision.Material and Methods: This comparative experimental study, utilizing retrospective data, also evaluated two state-of-the-art models, the Detection Transformer (DETR) and RetinaNet, for their performance. The models were trained and tested on the publicly available Categorized Digital Database for Low-Energy and Subtracted Contrast-Enhanced Spectral Mammography (CDD-CESM), which contains 1,982 mammograms with 3,720 annotated lesions of various types and sizes.Results: YOLO-v12 demonstrated excellent diagnostic accuracy (mean Average Precision at an IOU threshold of 0.5 (mAP50)=0.98; Intersection Over Union (IOU)=0.95), significantly outperforming contemporary models and older YOLO versions. Conclusion: The promising and robust results clearly underscore the remarkable potential of artificial intelligence technologies in effectively assisting radiologists with the early detection and diagnosis of breast cancer. These findings advocate for the implementation of YOLO-v12 in clinical mammography screening applications and suggest that future research should prioritize real-time diagnostic systems to further enhance breast cancer detection capabilities.

Dependence and correlation represent distinct statistical concepts. While there are methods to measure linear and nonlinear correlation between two variables, understanding the statistical dependence between them remains a topic of great interest. In this paper, we propose a heuristic, numerical, and algorithmic approach to estimate the dependence coefficient between two variables. With this approach, first, the X–Y scatter plot is transformed into a functional scatter plot using a procedure called “functionalizing.” Next, a novel concept called “successive triangles” is employed to estimate the dependence of Y on X. The proposed method offers several advantages; it is distribution-free, so it is suitable for both Gaussian and non-Gaussian numerical variables. Moreover, it can be used for both numerical and categorical (nominal) variables. This approach can be employed in other applications such as correlation measurement and also template matching for single-dimensional patterns. The presented method has been validated by both the simulated and clinical data with promising results.

From a psychological standpoint, noise can disturb human mental performance in the workplace. While loudness and sharpness of noise are usable factors for evaluating the subjective effects of noise exposure, there has been comparatively less focus on their influence on human reaction time and the dynamic changes in brain waves. Therefore, this study aimed to investigate the temporal changes in the theta beta ratio (TBR) and reaction time (RT) before and during exposure to different noise loudness and sharpness levels. Eight different loudness and sharpness levels of noise were generated and broadcast. Sixty-four men performed the Integrated Visual and Auditory (IVA) test both before and during exposure to the generated noise, and their brain waves were recorded simultaneously. The main step of the IVA test lasted approximately 12 min and consisted of 250 visual and 250 auditory stimuli presented randomly. The highest difference in TBR before and during exposure to various loudness levels was associated with 65 phon at 250 Hz and 75 phon at 4000 Hz, with values of 0.86 and 0.75, respectively. This difference equaled 0.69 for the 1.23 acum sharpness level at a frequency of 4000 Hz. In addition, the sharpness of 3.29 acum at a frequency of 8000 Hz led to the highest difference equaling 0.71 in TBR between before and during exposure to various sharpness levels. The most significant difference in RT was observed before and during exposure to lower loudness levels (65 and 75 phon), resulting in 405.75 and 347.5 ms, respectively. This highest difference was also observed for sharpness levels of 1.23 (391.4 ms) and 3.29 acum (427.4 ms) compared to other sharpness levels. Furthermore, attention deficit increased with longer exposure times to all noise levels. The results showed a statistically significant correlation between TBR and RT during exposure to all noises. Specifically, lower loudness at frequencies of 250 and 4000 Hz and higher sharpness at frequencies of 4000 and 8000 Hz had the most significant impact on the attention parameter. In psychoacoustics studies, it is important to consider both the frequency and duration of exposure in addition to assessing the loudness and sharpness level of noise.

Background: One of the main reasons for neonatal deaths is preterm delivery, and infants who have survived preterm birth (PB) are at risk of significant health complications. However, an effective method for reliable and accurate prediction of preterm labor has yet to be proposed.Objective: This study proposes an artificial neural network (ANN)-based approach for early prediction of PB, and consequently can hint physicians to start the treatment earlier, reducing the chance of morbidity and mortality in the infant.Material and Methods: This historical cohort study proposes a feed-forward ANN with 7 hidden neurons to predict PB. Thirteen risk factors of PB were collected from 300 pregnant women (150 with preterm delivery and 150 normal) as the ANN inputs from 2018 to 2019. From each group, 70%, 15%, and 15% of the subjects were randomly selected for training, validation, and testing of the model, respectively.Results: The ANN achieved an accuracy of 79.03% for the classification of the subjects into two classes normal and PB. Moreover, a sensitivity of 73.45% and specificity of 84.62% were obtained. The advantage of this approach is that the risk factors used for prediction did not require any lab test and were collected in a questionnaire. Conclusion: The efficacy of the proposed approach for the early identification of pregnant women, who are at high risk of preterm delivery, leads to necessary care and clinical interventions, applied during the pregnancy.

Bacterial infection is a life-threatening situation, and its rapid diagnosis is essential for treatment. Apart from medical applications, rapid identification of bacteria is vital in the food industry or the public health system. There are various bacterial identification techniques, including molecular-based methods, immunological approaches, and biosensor-based procedures. The most commonly used methods are culture-based methods, which are time-consuming. The objective of this study is to find a fingerprint of bacteria to identify them. Three strains of bacteria were selected, and seven different concentrations of each bacterium were prepared. The bacteria were then treated with two different molar concentrations of the fluorescent fluorophore, dichlorodihydrofluorescein diacetate for 30 minutes. Then, using the fluorescence mode of a multimode reader, the fluorescence emission of each bacterium is scanned twice during 60 minutes. Plotting the difference between two scans versus the bacteria concentration results in a unique fluorescence pattern for each bacterium. Observation of the redox state of bacteria, during 90 minutes, results in a fluorescence pattern that is clearly a fingerprint of different bacteria. This pattern is independent of fluorophore concentration. Mean Squares Errors (MSE) between the fluorescence patterns of similar bacteria is less than that of different bacteria, which shows the method can properly identify the bacteria. In this study, a new label-free method is developed to detect and identify different species of bacteria by measuring the redox activity and using the fluorescence fluorophore, dichlorodihydrofluorescein diacetate. This robust and low-cost method can properly identify the bacteria, uses only one excitation and emission wavelength, and can be simply implemented with current multimode plate readers.

Bachground:Noise is one of the most important harmful factors in the environment. There are limited studies on the effect of noise loudness on brain signals and attention. The main objective of this study was to investigate the relationship between exposure to different loudness levels with brain index, types of attention, and subjective evaluation. Methods:Four noises with different loudness levels were generated. Sixty-four male students participated in this study. Each subject performed the integrated visual and auditory continuous performance test (IVA-2) test before and during exposure to noise loudness signals while their electroencephalography was recorded. Finally, the alpha-to-gamma ratio (AGR), five types of attention, and the subjective evaluation results were examined. Results:During exposure to loudness levels, the AGR and types of attention decreased while the NASA-Tax Load Index (NASA-TLX) scores increased. The noise exposure at lower loudness levels (65 and 75 phon) leads to greater attention dysfunction than at higher loudness. The AGR was significantly changed during exposure to 65 and 75 phon and audio stimuli. This significant change was observed in exposure at all loudness levels except 85 phon and visual stimuli. The divided and sustained attention changed significantly during exposure to all loudness levels and visual stimuli. The AGR had a significant inverse correlation with the total score of NASA-TLX during noise exposure. Conclusions:These results can lead to the design of methods to control the psychological effects of noise at specific frequencies (250 and 4000 Hz) and can prevent non-auditory damage to human cognitive performance in industrial and urban environments.

Abstract The authors have requested that this preprint be removed from Research Square.

Background: Mammography is the most reliable and popular method in the clinical diagnosis of breast cancer. Calcifications are subtle lesions in mammograms that can be cancerous and difficult to detect for radiologists. Computer-aided detection (CAD) can help radiologists identify malignant lesions. Objectives: This study aimed to propose a deep learning based CAD system for detecting calcifications in mammograms. Patients and Methods: A total of 815 in-house mammograms were collected from 204 women undergoing screening mammography. Calcifications in the mammograms were annotated by specialists. Each mammogram was divided into patches of fixed size, and then, patches containing calcifications were extracted, along with the same number of normal patches. A ResNet-50 Convolutional Neural Network (CNN) was trained for classification of patches into normal and calcification groups using training data and then the performance of the trained CNN was tested with new test data. Results: The proposed patch learning approach (PLA) showed a classification accuracy of 96.7% in the binary classification of patches. Therefore, it could detect calcification regions in a given mammogram. The PLA achieved sensitivity and specificity of 96.7% and 96.7%, respectively, with an area under the curve of 98.8%. Conclusion: The present results highlighted the efficacy of the proposed PLA, especially for limited training data. Direct comparison with previous studies is not possible due to differences in datasets. Nevertheless, the PLA accuracy in detecting calcifications was higher than that of deep learning based CAD systems in previous studies. The effective performance of PLA may be attributed to the manual removal of uninformative patches, as they were not used in the training set.

Efficient storage and transmission of electromyogram (EMG) data are important for emerging applications such as telemedicine and big data, as a vital tool for further advancement of the field. However, due to limitations in internet speed and hardware resources, transmission and storage of EMG data are challenging. As a solution, this work proposes a new method for EMG data compression using deep convolutional autoencoders (CAE). Eight-channel EMG data from 10 subjects, and high-density EMG data from 18 subjects, were investigated for compression. The CAE architecture was designed to extract an abstract data representation that is heavily compressed, but from which the salient information for classification can be effectively reconstructed. The proposed method attained efficient compression; for CR = 1600, the average PRDN (percentage RMS difference normalized) was 31.5% and the wrist motions classification accuracy (CA) reduced roughly 5%. The CAE substantially outperformed the state-of-the-art high-efficiency video coding and a well-known wavelet-thresholding compression technique. Moreover, by reducing the bit-resolution of the CAE's compressed data from 24 bits to 6 bits, an additional 4-fold compression was achieved without significant degradation of the reconstruction performance. Furthermore, the CAE's inter-subject performance was promising; e.g., for CR = 1600, the PRDN for the inter-subject case was only 2.6% less than that of the within-subject performance. The powerful EMG compression performance with remarkable reconstruction results reflects the CAEs potential as an automatic end-to-end approach with the ability to learn the complete encoding and decoding process. Furthermore, the excellent inter-subject performance demonstrates the generalizability and usability of the proposed approach.

Background: The conventional procedure of skin-related disease detection is a visual inspection by a dermatologist or a primary care clinician, using a dermatoscope. The suspected patients with early signs of skin cancer are referred for biopsy and histopathological examination to ensure the correct diagnosis and the best treatment. Recent advancements in deep convolutional neural networks (CNNs) have achieved excellent performance in automated skin cancer classification with accuracy similar to that of dermatologists. However, such improvements are yet to bring about a clinically trusted and popular system for skin cancer detection. Objective: This study aimed to propose viable deep learning (DL) based method for the detection of skin cancer in lesion images, to help physicians in diagnosis. Material and Methods: In this analytical study, a novel DL based model was proposed, in which other than the lesion image, the patient’s data, including the anatomical site of the lesion, age, and gender were used as the model input to predict the type of the lesion. An Inception-ResNet-v2 CNN pretrained for object recognition was employed in the proposed model. Results: Based on the results, the proposed method achieved promising performance for various skin conditions, and also using the patient’s metadata in addition to the lesion image for classification improved the classification accuracy by at least 5% in all cases investigated. On a dataset of 57536 dermoscopic images, the proposed approach achieved an accuracy of 89.3%±1.1% in the discrimination of 4 major skin conditions and 94.5%±0.9% in the classification of benign vs. malignant lesions. Conclusion: The promising results highlight the efficacy of the proposed approach and indicate that the inclusion of the patient’s metadata with the lesion image can enhance the skin cancer detection performance.

Breast cancer is one of the most common types of cancer in women. Screening mammography is a low‑dose X‑ray examination of breasts, which is conducted to detect breast cancer at early stages when the cancerous tumor is too small to be felt as a lump. Screening mammography is conducted for women with no symptoms of breast cancer, for early detection of cancer when the cancer is most treatable and consequently greatly reduce the death rate from the breast cancer. Screening mammography should be performed every year for women age 45-54, and every two years for women age 55 and older who are in good health. A mammogram is read by a radiologist to diagnose cancer. To assist radiologists in reading mammograms, computer-aided detection (CAD) systems have been developed which can identify suspicious lesions on mammograms. CADs can improve the accuracy and confidence level of radiologists in decision making and have been approved by FDA for clinical use. Traditional CAD systems work based on conventional machine learning (ML) and image processing algorithms. With recent advances in software and hardware resources, a great breakthrough in deep learning (DL) algorithms was followed, which revolutionized various engineering areas including medical technologies. Recently, DL models have been applied in CAD systems in mammograms and achieved outstanding performance. In contrast to conventional ML, DL algorithms eliminate the need for the tedious task of human-designed feature engineering, as they are capable of learning useful features automatically from the raw data (mammogram). One of the most common DL frameworks is the convolutional neural network (CNN). To localize lesions in a mammogram, a CNN should be applied in region‑based algorithms such as R‑CNN, Fast R‑CNN, Faster R‑CNN, and YOLO. Proper training of a DL‑based CAD requires a large amount of annotated mammogram data, where cancerous lesions have been marked by an experienced radiologist. This highlights the importance of establishing a large, annotated mammogram dataset for the development of a reliable CAD system. This article provides a brief review of the state‑of‑the‑art techniques for DL‑based CAD in mammography.

Background: Skin cancer is one of the most common forms of cancer in the world and melanoma is the deadliest type of skin cancer. Both melanoma and melanocytic nevi begin in melanocytes (cells that produce melanin). However, melanocytic nevi are benign whereas melanoma is malignant. This work proposes a deep learning model for classification of these two lesions. Methods: In this analytic study, the database of HAM10000 (human against machine with 10000 training images) dermoscopy images, 1000 melanocytic nevi and 1000 melanoma images were employed, where in each category 900 images were selected randomly and were designated as the training set. The remaining 100 images in each category were considered as the test set. A deep learning convolutional neural network (CNN) was deployed with AlexNet (Krizhevsky et al., 2012) as a pretrained model. The network was trained with 1800 dermoscope images and subsequently was validated with 200 test images. The proposed method removes the need for cumbersome tasks of lesion segmentation and feature extraction. Instead, the CNN can automatically learn and extract useful features from the raw images. Therefore, no image preprocessing is required. Study was conducted at Shahid Beheshti University of Medical Sciences, Tehran, Iran from January to February, 2020. Results: The proposed model achieved an area under the receiver operating characteristic (ROC) curve of 0.98. Using a confidence score threshold of 0.5, a classification accuracy of 93%, sensitivity of 94%, and specificity of 92% was attained. The user can adjust the threshold to change the model performance according to preference. For example, if sensitivity is the main concern; i.e. false negative is to be avoided, then the threshold must be reduced to improve sensitivity at the cost of specificity. The ROC curve shows that to achieve sensitivity of 100%, specificity is decreased to 83%. Conclusion: The results show the strength of convolutional neural networks in melanoma detection in dermoscopy images. The proposed method can be deployed to help dermatologists in identifying melanoma. It can also be implemented for self diagnosis of photographs taken from skin lesions. This may facilitate early detection of melanoma, and hence substantially reduce the mortality chance of this dangerous malignancy.

Background: The most common types of non-melanoma skin cancer are basal cell carcinoma (BCC), and squamous cell carcinoma (SCC). AKIEC -Actinic keratoses (Solar keratoses) and intraepithelial carcinoma (Bowen’s disease)- are common non-invasive precursors of SCC, which may progress to invasive SCC, if left untreated. Due to the importance of early detection in cancer treatment, this study aimed to propose a computer-based model for identification non-melanoma malignancies. Methods: In this analytic study, 327 AKIEC, 513 BCC, and 840 benign keratosis images from human against machine with 10000 training dermoscopy images (HAM10000) were extracted. From each of these three types, 90% of the images were designated as the training set and the remaining images were considered as the test set. A deep learning convolutional neural network (CNN) was developed for skin cancer detection by using AlexNet (Krizhevsky, et al., 2012) as a pretrained network. First, the model was trained on the training images to discriminate between benign and malignant lesions. In comparison with conventional methods, the main advantage of the proposed approach is that it does not need cumbersome and time-consuming procedures of lesion segmentation and feature extraction. This is because CNNs have the capability of learning useful features from the raw images. Once the system was trained, it was validated with test data to assess the performance. Study was carried out at Shahid Beheshti University of Medical Sciences, Tehran, Iran, in January and February, 2020. Results: The proposed deep learning network achieved an AUC (area under the ROC curve) of 0.97. Using a confidence score threshold of 0.5, a classification accuracy of 90% was attained in the classification of images into malignant and benign lesions. Moreover, a sensitivity of 94% and specificity of 86% were obtained. It should be noted that the user can change the threshold to adjust the model performance based on preference. For example, reducing the threshold increase sensitivity while decreasing specificity. Conclusion: The results highlight the efficacy of deep learning models in detecting non-melanoma skin cancer. This approach can be employed in computer-aided detection systems to assist dermatologists in identification of malignant lesions.

BACKGROUND:Myoelectric control is a method of decoding the motor intent from the electromyogram (EMG) data and using the estimated intent to control prostheses and robots. This work investigates estimation of the wrist kinematics from EMG signals using polynomial models. Due to their low complexity, polynomial models are potentially the perfect choice for EMG-kinematics modeling.METHODS:Ten ablebodied individuals participated in this study, where the EMG signals from the forearm and the wrist kinematics from the contralateral wrist were measured during mirrored contractions. Two sets of EMG features were employed including the time domain (TD) set, and TD features along with autoregressive coefficients (TDAR). Polynomial models of order 1 to 4 were applied to map the EMG signals to the wrist motions. The performance was directly compared to that of a multilayer perceptron (MLP) neural network.RESULTS:The estimation accuracy of the wrist kinematics improved with increasing the order of the model, but saturated at the 4th order. When using the TD set, the MLP significantly outperformed all polynomial models. However, when using the TDAR set, the polynomial models' performance improved so that the 4th order model performance was not significantly different than that of the MLP in two DoFs, although it was lower than MLP in one DoF.CONCLUSION:These results indicate that polynomial models are not as effective as more complex models such as neural networks, in learning the highly nonlinear mapping between the EMG data and motion intent. However, using a sufficiently high number of various EMG features, would reduce the mapping nonlinearities, and thereby may increase the polynomial models' performance to levels similar to those of complex black box models.

An important barrier to commercialization of pattern recognition myoelectric control of prostheses is the lack of robustness to confounding factors such as electrode shift, skin impedance variations, and learning effects. To overcome this challenge, a novel supervised adaptation approach based on transfer learning (TL) with convolutional neural networks (CNNs) is proposed which requires only a short training session (a few seconds for each class) to recalibrate the system. TL is proposed as a solution to the problem of insufficient calibration data due to short training times for both classification and regression-based control schemes. This approach was validated for electrode shift of roughly 2.5cm with 13 able-bodied subjects to estimate individual and combined wrist motions. With this method, the original CNN (trained before the shift) was fine-tuned with the calibration data from after shifting. The results show that the proposed technique outperforms training a CNN from scratch (random initialization of weights) or a support vector machine (SVM) using the minimal calibration data. Moreover, it demonstrates superior performance than previous LDA and QDA-based adaptation approaches. As the outcomes confirm, the proposed CNN TL method provides a practical solution for adaptation to external factors, improving the robustness of electromyogram (EMG) pattern recognition systems.

The polarimetry imaging technique has provided a powerful tool for discriminating normal from cancerous tissues. In this paper, based on the backscattering Mueller matrix imaging of prostate bulk tissues, (received immediately after surgery without any further processing), we have extracted the characteristic features of the Mueller matrix images. In order to provide a quantitative and more accurate comparison, three different methods have been used; the Mueller matrix polar decomposition (MMPD), the Mueller matrix transformation (MMT) and the frequency distribution histograms (FDHs) and their central moment parameters. Comparing different tissues, the results of our study indicate that these methods provide the indicators for the characteristics of the microstructural features of the tissues. The indicators have the potential to distinguish between cancerous and healthy tissues. Determining the polarimetric characteristics of the tissue immediately after surgery and prior to the pathology, and the potential possibility of this technique to be used in vivo as an optical biopsy technique, can significantly reduce the cost and time of diagnosis of cancer.

Objective. Deep learning models can learn representations of data that extract useful information in order to perform prediction without feature engineering. In this paper, an electromyography (EMG) control scheme with a regression convolutional neural network (CNN) is proposed as a substitute of conventional regression models that use purposefully designed features. Approach. The usability of the regression CNN model is validated for the first time, using an online Fitts' law style test with both individual and simultaneous wrist motions. Results were compared to that of a support vector regression-based scheme with a group of widely used extracted features. Main results. In spite of the proven efficiency of these well-known features, the CNN-based system outperformed the support vector machine (SVM) based scheme in throughput, due to higher regression accuracies especially with high EMG amplitudes. Significance. These results indicate that the CNN model can extract underlying motor control information from EMG signals during single and multiple degree-of-freedom (DoF) tasks. The advantage of regression CNN over classification CNN (studied previously) is that it allows independent and simultaneous control of motions.

Background: Deep learning has revolutionized artificial intelligence and has transformed many fields. It allows processing high-dimensional data (such as signals or images) without the need for feature engineering. The aim of this research is to develop a deep learning-based system to decode motor intent from electromyogram (EMG) signals. Methods: A myoelectric system based on convolutional neural networks (CNN) is proposed, as an alternative to conventional classification methods that depend on feature engineering. The proposed model was validated with 10 able-bodied subjects during single and combined wrist motions. Eight EMG channels were recorded using eight pairs of surface electrodes attached around the subject’s dominant forearm. The raw EMG data from windows of 167ms (200 samples) in 8 channels were arranged as 200×8 matrices. For each subject, a CNN was trained using the EMG matrices as the input and the corresponding motion classes as the target. The resulting model was tested using a 4-fold cross-validation. The performance of the proposed approach was compared to that of a standard SVM-based model that used a set of time-domain (TD) features including mean absolute value, zero crossings, slope sign changes, waveform length, and mean frequency. Results: In spite of the proven performance and popularity of the TD features, no significant difference (P=0.19) was found between the classification accuracies of the two methods. The advantage of the proposed model is that it does not need manual extraction of features, as the CNN can automatically learn and extract required representations from the EMG data. Conclusion: These results indicate the capacity of CNNs to learn and extract rich and complex information from biological signals. Because both amplitude and frequency of EMG increases with increasing muscle force, both temporal and spectral characteristics of EMG are needed for efficient estimation of motor intent. The TD set, also includes these types of features. The high performance of the CNN model shows its capability to learn temporal and spectral representations from raw EMG data.

The evolution of deep learning techniques has been transformative as they have allowed complex mappings to be trained between control inputs and outputs without the need for feature engineering. In this work, a myoelectric control system based on convolutional neural networks (CNN) is proposed as a possible alternative to traditional approaches that rely on specifically designed features. This CNN-based system is validated using a real-time Fitts’ law style target acquisition test requiring single and combined wrist motions. The performance of the proposed system is then compared to that of a standard support vector machine (SVM) based myoelectric system using a set of time-domain features. Despite the prevalence and demonstrated performance of these well-known features, no significant difference (p>0.05) was found between the two methods for any of the computed control metrics. This demonstrates the potential for automated learning approaches to extract complex and rich information from stochastic biological signals. This first evaluation of the usability of a CNN in a real-time myoelectric control environment provides a basis for further exploration.