Search

Introduction: Several types of cancer can be detected early through thermography, which uses thermal profiles to image tissues in recent years, thermography has gained increasing attention due to its non-invasive and radiation-free nature. There is a growing need for thermographic images of breast cancer lesions in different nationalities and ages to develop this technique, however. This study aims to introduce a dataset of breast thermograms. Material and Methods: In this study, thermographic images of breast cancer from Iranian samples were prepared and confirmed due to the limited number of breast thermogram databases. The prepared database was tested using artificial intelligence and another well-known DMR database (Database for Mastology Research) in this study to determine its reliability. Results: A variety of deep learning architectures and transfer learning are used to evaluate these databases for accuracy, sensitivity, speed, training compliance, and validation compliance. According to best-fitted structures for both types of databases, the database obtained from this study has a quality comparable to the DMR reference database, with minimum accuracy, sensitivity, specificity, precision, and F-score of 80%, 86%, 86%, 88%, and 87%, respectively. Conclusion: Using thermography as a method of early breast screening is demonstrated to be effective. In comparison to DMR, the lower statistics of the proposed database (between 2 and 7 percent) indicates that more diverse breast thermograms should be captured in conjunction with improvements to imaging equipment as well as adherence to thermography recording protocols in order to improve the reliability and efficiency of the database. [ABSTRACT FROM AUTHOR]

A sudden fall accident is the main concern for the elderly and disabled people. Automatic detection of the falls from video sequences is an assistive technology for surveillance systems. In this study, a three-stage framework was presented and implemented based on the combination of the data from multiple cameras to address the challenges of occlusion and visibility. In the first stage, the number of used cameras was specified. In the second stage, each camera was decided locally based on its data about the fall incident. In the third and final stage, the aggregation function was used to combine the single camera's decision considering the coverage rate coefficient of the used cameras. Experiments on the multiple-camera fall dataset demonstrated that our method is comparable to other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Introduction: Credible research shows that in many cases, infertility is caused by problems with men's sperm. Therefore, an accurate sperm analysis is a necessity to solve the problem of infertile couples. However, the first step in this analysis is to separate the sperm from the semen, which may not be accurate due to the poor contrast of the captured microscopy images. Material and Methods: Curvelet transform has been introduced in recent years as an effective tool for object detection in the image processing domain. The most significant advantage of this transform is that it maps the raw image to a new space in which the features are sparser and parallelly more directional. Based on this fact, in this study, this approach is used to more effectively detect sperm in microscopic images. However, intelligent adjustment of the parameters of this mapping plays an important role in strengthening the weaker edges, and therefore in this article, a new method for optimizing the mapping parameters in order to achieve better separation of sperms from the background of the semen image is also proposed and examined. Results: The comparison of the results obtained from examining the proposed method versus the results of the state of art methods was performed by using the two main criteria including sperm detection rate and false detection rate (i.e., false positives). This comparison clearly indicated the effectiveness of the proposed idea in distinguishing sperms from semen background. When the basis of performance evaluation is based on not detecting even a single false particle, it is observed that the correct sperm detection rate in the proposed method is between 4 and 17 percent higher than alternative methods. However, the false detection parameter itself shows an improvement of 33% to 3% in the proposed method compared to the weakest to the best among alternatives. Conclusion: Investigating the ROC curve which has been obtained from several examinations showed the effectiveness of the proposed idea over its alternatives either in correct detection of sperms or elimination of false objects. Therefore, the obtained results may lead us to the conclusion that the curvelet transform may be utilized as an effective solution for detecting sperms in low contrast microscopic images of human semen. [ABSTRACT FROM AUTHOR]

Objectives: Automated Auditory Brainstem Responses (ABR) peak detection is a novel technique to facilitate the measurement of neural synchrony along the auditory pathway through the brainstem. Analyzing the location of the peaks in these signals and the time interval between them may be utilized either for analyzing the hearing process or detecting peripheral and central lesions in the human hearing system. Methods: In this paper, model-based signal processing is proposed to estimate the effective parameters of ABR signals. In this process, the biological parameters of the signal are assessed by utilizing a Finite Impulse Response (FIR) adaptive filter in which its adaptation procedure is performed based on the correntropy concept. The proposed method is applied on a set of ABR signals recorded in response to three stimuli of /da/, /ba/, and /ga/, and then its performances are compared with an existing state-of-the-art technique. Results: The results show that the proposed method can significantly increase the accuracy of estimating the parameters in stable stimulations (/da/, /ba/) for major positive and negative peaks. This improvement is more significant (up to 2-3 times) for /ba/ stimulus and especially in major positive peaks. However, in other peaks, the improvements also occurred in smaller amounts. However, for unstable stimuli (/ga/), no significant improvement was achieved. Discussion: Increasing the accuracy performance of the proposed method for detecting the stable stimuli (while its performance remains unchanged) for detecting unstable stimuli indicates its effectiveness in automated clinical analysis of ABR signals. [ABSTRACT FROM AUTHOR]

Background: Deep neural networks have been widely used in detection of P300 signal in Brain Machine Interface (BMI) systems which are rely on Event-Related Potentials (ERPs) (i.e. P300 signals). Such networks have high curvature variation in their error surface hampering their favorable performance. Therefore, the variations in curvature of the error surface must be minimized to improve the performance of these networks in detecting P300 signals. Objective: The aim of this paper is to introduce a method for minimizing the curvature of the error surface during training Convolutional Neural Network (CNN). The curvature variation of the error surface is highly dependent on model parameters of deep neural network; therefore, we try to minimize this curvature by optimizing the model parameters. Material and Methods: In this experimental study an attempt is made to tune the CNN parameters affecting the curvature of its error surface in order to obtain the best possible learning. For achieving this goal, Genetic Algorithm is utilized to optimize the above parameters in order to minimize the curvature variations. Results: The performance of the proposed algorithm was evaluated on EPFL dataset. The obtained results demonstrated that the proposed method detected the P300 signals with maximally 98.91% classification accuracy and 98.54% True Positive Ratio (TPR). Conclusion: The obtained results showed that using genetic algorithm for minimizing curvature of the error surface in CNN increased its accuracy in parallel with decreasing the variance of the results. Consequently, it may be concluded that the proposed method has considerable potential to be used as P300 detection module in BMI applications. [ABSTRACT FROM AUTHOR]

Introduction: Recent studies have acknowledged the potential of convolutional neural networks (CNNs) in distinguishing healthy and morbid samples by using heart sound analyses. Unfortunately the performance of CNNs is highly dependent on the filtering procedure which is applied to signal in their convolutional layer. The present study aimed to address this problem by applying filter bank learning concept in CNNs. Material and Methods: In proposed method, the filter bank of CNN is updated based on a cross-entropy minimization rule to extract higher-level features from spectral characteristics of the heart sound signal. The deeper level of the extracted features in parallel with their spectral-based nature leads to better discrimination between healthy and morbid heart sounds. The proposed method was applied to three different heart sound datasets of PASCAL-A, PASCAL-B, and Kaggle, including normal and abnormal categories. Results: The proposed method obtained a true positive rate (TPR) between minimally 86% and maximally 96% (if FPR=0%) among all the examined datasets. In addition, the false-positive rate (FPR) was obtained as 7-8% (if TPR=100%) among the mentioned datasets. Finally, the accuracy was achieved in the range of 93- 98% when the FPR was 0% and within the range of 96-96.5% when the TRP was 100%. Conclusion: Increased TPR in the proposed method (96% for the proposed method vs. 87% for CNN) in parallel with a decrease in its FPR (7% for the proposed method vs. 10% for CNN) showed the proposed method's superiority against its well-known alternative in automated self-assessment of the heart. [ABSTRACT FROM AUTHOR]

Vertex colouring is among the most important problems in graph theory which has been widely applied across different real-world problems. In vertex colouring problem (VCP), the goal is to assign a distinct colour to each vertex of the graph in such a way that no two adjacent vertices have the same colour. This paper presents a fuzzy irregular cellular automaton (FICA) for finding a near-optimal solution for the VCP. FICA is an extension fuzzy cellular automaton (FCA) in which the cells of the automaton can be arranged in an irregular structure. The aim of the proposed method is to reap the benefits of both FCA and irregular cellular automata while minimising their drawbacks. To evaluate the proposed method, various computer simulations have been conducted on a variety of graphs. The results suggest that the proposed method is able to achieve better results in terms of the minimum number of required colours and the execution time of the algorithm, compared to other peer algorithms. [ABSTRACT FROM AUTHOR]

Introduction: In recent years, modern microscopic imaging in parallel with digital image processing techniques, have facilitated computerized semen analysis. However, in these methods, distinguishing sperms from other semen particles can be hampered by low contrast of microscopic images and the possibility of neighboring sperms touching each other. Materials and Methods: This article introduced a new method based on combination of Zernike moments and spatial processing in sperm detection. In the first step, Zernike moments were estimated due to their rotation and noise-resistant nature to mark pixels with some chance of belonging to sperms. In the second step, pruning was executed considering the connectivity of candidates and using morphological processing, to extract sperms. The proposed algorithm was examined on microscopic images exhibiting several sperms with different morphologies. Results: The obtained results showed the ability of the proposed method in sperm detection, such that it could detect 85% of the sperms without any false detection. In a more pragmatic situation, where false detection rate was 5%, the detection rate of the proposed algorithm increased to 94%. Conclusion: Comparing the proposed method with watershed segmentation algorithm (WSA) and morphological contour synthesis (MCS) revealed the superiority of the proposed method to its alternatives in such a way that it detected sperms at least 3% and 13% better than WSA and MCS, respectively, without any false detection. Furthermore, the rate of false detections in the proposed algorithm was at least 4% and 14% better than its alternatives. [ABSTRACT FROM AUTHOR]

Background: P300 signal detection is an essential problem in many fields of Brain-Computer Interface (BCI) systems. Although deep neural networks have almost ubiquitously used in P300 detection, in such networks, increasing the number of dimensions leads to growth ratio of saddle points to local minimums. This phenomenon results in slow convergence in deep neural network. Hyperparameter tuning is one of the approaches in deep learning, which leads to fast convergence because of its ability to find better local minimums. In this paper, a new adaptive hyperparameter tuning method is proposed to improve training of Convolutional Neural Networks (CNNs). Methods: The aim of this paper is to introduce a novel method to improve the performance of deep neural networks in P300 signal detection. To reach this purpose, the proposed method transferred the non-convex error function of CNN) into Lagranging paradigm, then, Newton and dual active set techniques are utilized for hyperparameter tuning in order to minimize error of objective function in high dimensional space of CNN. Results: The proposed method was implemented on MATLAB 2017 package and its performance was evaluated on dataset of Ecole Polytechnique Fédérale de Lausanne (EPFL) BCI group. The obtained results depicted that the proposed method detected the P300 signals with 95.34% classification accuracy in parallel with high True Positive Rate (i.e., 92.9%) and low False Positive Rate (i.e., 0.77%). Conclusions: To estimate the performance of the proposed method, the achieved results were compared with the results of Naive Hyperparameter (NHP) tuning method. The comparisons depicted the superiority of the proposed method against its alternative, in such way that the best accuracy by using the proposed method was 6.44%, better than the accuracy of the alternative method. [ABSTRACT FROM AUTHOR]

Background: A challenging step in breast thermography for detecting cancer is to extract the breast region and remove the rest of the tissues in order to increase the accuracy of the final decision. Unfortunately, the performance of this procedure is hampered by low contrast of thermograms. Objective: The aim of this article is promoting fuzzy concept in order to local modeling of the pixels in a breast thermogram. By applying the fuzzy rules on contrast and lowpass filtered versions of thermogram, each pixel of the image is scored as either low, medium or high. Then the extracted edges of the fuzzy processed thermogram indicate the region of breast. Material and Methods: The proposed algorithm was applied to a DMR-IR well known breast thermography dataset including thermal images which have been captured with two static and dynamic protocols. The method implemented in Tensor- flow framework as a machine learning library in Python in Google Colab (i.e., a GPU framework provided by Google) testbed. Results: The proposed local fuzzy scheme reached to MSE, PSNR and Jaccard similarity coefficient of 0.0015, 29.08 and 0.9399 in detecting breast. These values show significant improvement in performance compared to results of basic fuzzy method which led to 0.0038, 24.77 and 0.90, for above parameters respectively. Conclusion: Improving the performance criteria in proposed local fuzzy method shows that it may be considered as a candidate for use in future researches related to the detection of breast cancer in thermography systems. [ABSTRACT FROM AUTHOR]

Background: In last decade, in parallel with the expansion of the use of IOT in the field of health, its use in recognizing activities of the elderly who live alone has also been proposed. One of the main challenges of this technology is similarity of the signals which are captured from different activities. Objective: The aim of this article is utilizing Bi-directional Long Short-Term Memory (i.e., Bi-LSTM) which makes use of two independent temporal modeling together which leads the system to have both backward and forward information about the human activity signals, therefore may promote recognizing of human activities. Material and Methods: The tests were performed on Wireless Sensor Data Mining (WISDM) dataset including signals recorded from walking, jogging, upstairs, downstairs, standing and sitting activities. The basic LSTM were applied on above signals in parallel with Bi-LSTM to compare their accuracies. Results: The obtained results showed that Bi-LSTM reached to average accuracy of 93.54% in distinguishing several activities. The obtained accuracy by using basic LSTM were 89.46%. The highest increase in accuracy has been achieved for two critical activities upstairs and downstairs, whose accuracy level by LSTM was 77.80% and 83.18%, while using Bi-LSTM increased these values up to 79.20% and 98.42% respectively. Conclusion: Increasing of the recognition rate by using the proposed method specially for critical activities shows its effectiveness in IOT-based elderly health monitoring systems which may promote the level of public health. [ABSTRACT FROM AUTHOR]

Shape and movement features of sperms are important parameters for infertility study and treatment. In this article, a new method is introduced for characterization sperms in microscopic videos. In this method, first a hypothesis framework is defined to distinguish sperms from other particles in captured video. Then decision about each hypothesis is done in following steps: Selecting some primary regions as candidates for sperms by watershed-based segmentation, pruning of some false candidates during successive frames using graph theory concept and finally confirming correct sperms by using their movement trajectories. Performance of the proposed method is evaluated on real captured images belongs to semen with high density of sperms. The obtained results show the proposed method may detect 97% of sperms in presence of 5% false detections and track 91% of moving sperms. Furthermore, it can be shown that better characterization of sperms in proposed algorithm doesn't lead to extracting more false sperms compared to some present approaches. [ABSTRACT FROM AUTHOR]

Introduction The morphologic features of human sperms are key indicators for monitoring fertility problems in men. Therefore, automated analyzing methods via microscopic videos have become the most favorite policy in infertility treatment during the last decades. Materials and Methods In the proposed method, firstly a hypothesis testing framework was defined to distinguish sperms from background. Then, some regions were selected as candidates by minimization of the information distance between the original and processed images. Finally, the correct sperms were extracted from candidates using a watershed-based algorithm. Results The proposed, Watershed Segmentation Algorithm (WSA), Multi Structure Element Segmentation (MSES) and Dynamic Threshold Segmentation (DTS) algorithms achieve true positive rates of 96%, 84%, 81%, and 70%, respectively, versus typically 3% of false positive rate in semen specimens with high density of sperms. The true positive rates of 87%, 69%, 66%, and 52%, respectively, at the same false positive rate were obtained for the semen specimens with high density of sperms. Conclusion Results show that false positive rates of the proposed algorithm were at least 8% (in the first scenario) and 32% (in the second scenario) better than other methods considering the minimum acceptable true positive rate of 90%. Furthermore, it has been shown that the proposed algorithm extracted sperms at least 12% (in the first scenario) and 18% (in the second scenario) better than other methods in presence of a typically low false positive rate equal to 3%. [ABSTRACT FROM AUTHOR]

In this article, a new method is introduced for joint delay and Doppler estimation in ambiguity function (AF)-based radars. In this method, first each cell of AF is considered as a random variable and then a stochastic process is estimated for each cell based on its values during consecutive radar scans. In the second step, the AF is divided to high probability target and high probability clutter zones using parameters of the estimated stochastic processes. Finally, exact values of delay and Doppler of radar targets are extracted and localised from the divided AF by employing spatial processing techniques. Performance of the proposed method is evaluated in two different scenarios, corresponding to high and low speed targets, respectively. The obtained results show the greater ability of the suggested method in detection of the above types of targets compared to the present approaches. Furthermore, it can be shown that the proposed method causes more considerable improvement in the detection of low speed targets than high speed targets compared to available methods. [ABSTRACT FROM PUBLISHER]

Introduction: Studying motility of biological objects is an important parameter in many biomedical processes. Therefore, automated analyzing methods via microscopic videos are becoming an important step in recent researches. Materials and Methods: In the proposed method of this article, a hypothesis testing function is defined to separate biological particles from artifact and noise in captured video. Then, a decision about each hypothesis is made in the following steps: selecting primary candidates using stochastic modeling, pruning false candidates using graph theory, and confirming remained particles by Kalman filtering. Results: Performance of the proposed method is evaluated on real videos containing low and high densities of live Listeria particles. The results show that in the first scenario, the proposed and MD algorithms detect 95% and 65% of particles in presence of 2% and 44% false detections, respectively. In the second scenario, the proposed and MD algorithms detect 91% and 55% of particles in presence of 14% and 45% false detections, respectively. Conclusion: In the first scenario, the proposed algorithm detects and tracks particles typically 30% and 31% better than MD. Moreover, its false detected particles and trajectories are 42% and 27% less than MD. In the second scenario, the proposed method detects and tracks particles typically 36% and 38% better than MD, also its false detected particles and trajectories are 31% and 18% less than MD. Consequently, better characterization of particles in proposed algorithm not only does not lead to extracting more false particles and trajectories but also decreases the rate of false characterized particles and trajectories compared with existing algorithms. [ABSTRACT FROM AUTHOR]

Background: In Persian medicine (PM), measuring the wrist pulse is one of the main methods for determining a person's health status and temperament. One problem that can arise is the dependence of the diagnosis on the physician's interpretation of pulse wave features. Perhaps, this is one reason why this method has yet to be combined with modern medical methods. This paper addresses this concern and outlines a system for measuring pulse signals based on PM., Methods: A system that uses data from a customized device that logs the pulse wave on the wrist was designed and clinically implemented based on PM. Seven convolutional neural networks (CNNs) have been used for classification., Results: The pulse wave features of 34 participants were assessed by a specialist based on PM principles. Pulse taking was done on the wrist in the supine position (named Malmas in PM) under the supervision of the physician. Seven CNNs were implemented for each participant's pulse characteristic (pace, rate, vessel elasticity, strength, width, length, and height) assessment, and then, each participant was classified into three classes., Conclusion: It appears that the design and construction of a customized device combined with the deep learning algorithm can measure the pulse wave features according to PM and it can increase the reliability and repeatability of the diagnostic results based on PM., Competing Interests: There are no conflicts of interest., (Copyright: © 2022 Journal of Medical Signals & Sensors.)

Background: Recently, magnetic resonance imaging (MRI) has become a useful tool for the early detection of heart failure. A vital step of this process is a valid measurement of the left ventricle's properties, which seriously depends on the accurate segmentation of the heart in captured images. Although various schemes have been tested for this segmentation so far, the latest proposed methods have used the concept of deep learning to estimate the range of the left ventricle in cardiac MRI images. While deep learning methods can lead to better results than their classical alternatives, but unfortunately, the gradient vanishing and exploding problems may hamper their efficiency for the accurate segmentation of the left ventricle in MRI heart images., Methods: In this article, a new concept called residual learning is utilized to improve the performance of deep learning schemes against gradient vanishing problems. For this purpose, the Residual Network of Residual Network (i.e., Residual of Residual) substructure is utilized inside the main deep learning architecture (e.g., Unet), which provides more significant detection indexes., Results and Conclusion: The proposed method's performances and its alternatives were evaluated on Sunnybrook Cardiac Data as a reliable dataset in the left ventricle segmentation. The results show that the detection parameters are improved at least by 5%, 3.5%, 8.1%, and 11.4% compared to its deep alternatives in terms of Jaccard, Dice, precision, and false-positive rate indexes, respectively. These improvements were made when the recall parameter was reduced to a negligible value (i.e., approximately 1%). Overall, the proposed method can be used as a suitable tool for more accurate detection of the left ventricle in MRI images., Competing Interests: There are no conflicts of interest., (Copyright: © 2021 Journal of Medical Signals & Sensors.)