The main part of the eye is the retina covering the entire back section of the eye. Eye disease is one of the most important cause of disability and even death in developed countries as well as in developing countries. Disorders created in the retina that occur due to special diseases can be detected by specific retinal images. Studying the variations in retinal photos in a special time could help physicians to diagnose the associated diseases. In this paper, the detection of blood veins in retina photos was investigated. For this purpose, first a new method is proposed to promote the quality of retina photos by combining the histogram adjustment and gray level grouping. We use the feature vector to classify the pixels. Next, a method for classifying the images based on the feature extraction vector is required. The use of neural networks is one of the best and most widely used methods of machine learning for classification. We used a 3-layer Perceptron to classify pixels. تفاصيل المقالة

The production of additional cells often forms a mass of tissue that is referred to as a tumor. Tumors can disrupt the proper functioning of the brain and even lead to the patients' death. One of the non-invasive diagnostic methods for this disease is Magnetic Resonance Imaging (MRI). The development of an automated or semi-automatic diagnostic system is required by the computer in medical treatments. Several algorithms have been used to detect a tumor, each with its own advantages and disadvantages. In the present study, an automatic method has been developed by the combination of new methods in order to find the exact area of the tumor in the MRI image. This algorithm is based on super pixel and RVM classification. The algorithm used in the super pixel method is the SLIC algorithm, which calculates for each super pixel 13 statistical characteristics and severity. Finally, an educational method introduced from the RVM classification algorithm that can detect the tumor portion from non-tumor in each brain MRI image. BRATS2012 dataset and FLAIR weights have been utilized in this study The results are compared with the results of the BRATS2012 data and The overlap coefficients of Dice, BF score, and Jaccard were 0.898, 0.697 and 0.754, respectively. تفاصيل المقالة

Today, biometric recognition (based on biological signs) is a common and reliable method for recognizing and identity confirmation based on their behavioral and physiological characteristics. Physiological characteristics are consistent with the physical characteristics of individuals such as fingerprints, iris pattern, facial features, and the like. This type of property often does not change without external exertion. Behavioral characteristics such as signature, spoken pattern and iris are also a scale for identification and identity confirmation. In this study, using the wavelet method, the efficiency of human identification was increased by 75%. تفاصيل المقالة

Face recognition has been one of the most widely used sub-disciplines of machine learning for so many years. Face detection has been employed as an effective method in a wide range of applications such as surveillance systems and Forensic pathology in the area of machine vision. However, the accuracy of face detection has dramatically declined over the past decade due to wide-ranging challenges such as face detection with changes in face angle, the density of the crowds in an image, quality of light, etc which require special attention of researchers in response to these challenges. In the present study, a new sustainable approach to light changes for face detection based on local features is employed. In this method, the local binary pattern is extracted from face images and Principal Component Analysis is utilized to reduce the feature vectors’ dimension by the descriptor. Eventually, the features are classified using Ada Boost. Tests done on the images on the web show that face recognition accuracy is 100% in the low density crowd, 96% in the high-density crowd and proper light conditions, and 90% in the high-density crowd and poor light conditions. تفاصيل المقالة

By expanding the use of digital images in various areas of everyday life, such as medicine, identification, satellite imagery, and even personal cameras and machine vision, it is felt more effective in applying quality improvements to the images used. The low-quality images in the machine's vision can expose the efficacy of later processing, such as feature extraction, classification, and pattern recognition. In this thesis, a new method for improving the quality of images based on the extraction of Godin’s combined feature and model has been proposed. Based on the fact that each homogeneous region in the image has a Gaussian distribution histogram, this distribution can be divided into smaller histograms. For the histogram division efficacy, the image is transmitted from the RGB space to the HSV space and the histogram division is applied to the severity region, and the histogram is applied to each sub Histogram based on the statistical characteristics, and the image. Improved results are returned to the RGB color space. Several qualitative and quantitative criteria have been used to evaluate the proposed method. Qualitative comparison results show improved image quality compared to histogram equivalence methods and linear contrast traction. Quantitative evaluation criteria, such as entropy and spatial frequency, as well as signal to noise ratio, and peak signal to noise ratio, are generally proposed for superiority of the method. تفاصيل المقالة

All human biological traits are unique as biometrics, such as fingerprint, palm, iris, palm veins, finger veins and other biometrics. Using these biometrics has always been challenging. One of the challenges in biometrics is physical injuries. Finger vein biometrics is one of the characteristics that is most resistant to physical injuries. Numerous algorithms for authentication have been proposed with the help of this biometrics, which have weaknesses such as high computational complexity and low identification accuracy. In this paper, a new method in identification based on maximum curvature algorithm and morphological operators is proposed. The maximum curvature algorithm extracts image properties using a set of operations based on image returns. This process has been enhanced in the proposed method with morphological operators. What distinguishes the proposed method from other methods is that this algorithm is very accurate in distinguishing images which are similar but belonging to different classes. The proposed method, in addition to having a reasonable computational complexity, has been able to record very good identification accuracy in the challenge of low image quality. The identification accuracy of the proposed method is 97.5%, which compared to other methods has been able to improve more than 3%. Also, the identification speed of the proposed method is 0.84 seconds, which is very fast in its kind. تفاصيل المقالة

Brain is the most important organ in human body. All of our memories are saved on brain. Brain activity is analyzed by electroencephalogram signals (EEG). Brain activity and memory signal represent brain activity that can be recorded in different brain regions. Electroencephalography signal analysis can provide complete and comprehensive information about brain activity. Working of brain memory as an activity is analyzed by EEG. The main purpose of this research is to review the types of memory and especially working memory in humans by processing of EEG. In this regard, memory and types of memory have been discussed at the beginning. Then the brain activity and memory signal, recording method, electrode placement, brain potentials are discussed. Then, the complexity of the brain activity and memory signal in memory has been investigated. Then, memory and its relationship with brain activity and memory signal have been discussed. Areas affected by memory are expressed in the brain. After the researches about brain activity and memory signals in memory, it has been investigated. تفاصيل المقالة

Stomach cancer destroys the tissues of the digestive system. This cancer is one of the deadliest diseases. Endoscopic imaging is used to diagnose cancer. In endoscopy, the diagnosis of gastric cancer is difficult due to the similarity of the tissues, the low contrast of the image and the background. In order to overcome these problems, discrete wavelet transform has been used to detect stomach cancer. In the proposed method, there are data registration, data preprocessing, feature extraction, dimensionality reduction, and classification. The features are extracted with the help of discrete wavelet transform and then dimension reduction is done with the help of principal component analysis. The proposed approach was evaluated on datasets collected from five classes, including gastritis, ulcer, esophagitis , bleeding, and healthy. Jungle Tassafi has a value above 99% in all evaluation criteria, which represents the advantages of this category. The results of this research show that this method is accurate and reliable in diagnosis. تفاصيل المقالة

Breast cancer is the most common type of cancer among women worldwide. If diagnosed by a doctor in the early stages, it can save the patient's life. Ultrasound imaging is one of the most widely used diagnostic tools for diagnosing and classifying breast abnormalities. However, accurate segmentation of the ultrasound image is a challenging problem due to the artifacts created on the ultrasound image. Although deep learning-based methods have been able to overcome some of these challenges, the accuracy of tumor region detection in this image is still low. In this paper, we have proposed approaches for breast ultrasound image segmentation based on auto-encoder deep neural network. The proposed method has two parts. The classification section to determine the image with cancerous tissue and the tumor segmentation section to segment the desired area. which will be shown in the network output of the encoder itself. The proposed method has been evaluated qualitatively and quantitatively. The superiority of the proposed method with accuracy and dice criteria is 89 and 90 percent, respectively which shows the effectiveness of this method in diagnosis. تفاصيل المقالة

Face recognition is a widely used identification method in the machine learning field because face biometrics are distinctive enough for detection and have more accessibility compared to other biometrics. Despite their merits, face biometrics have various challenges. Mainly, these challenges are divided into local and global categories. Local challenges can be addressed using sustainable methods against change while global challenges such as illumination challenges require powerful pre-processing methods. Therefore, in this study, a sustainable method against light changes has been proposed. In this method, two stages of the Difference of Gaussian have been utilized for the illumination normalization. Then, the features of the normalized image are extracted using Histogram of Oriented Gradient (HOG) and the feature vectors are classified using 3 k-nearest neighbor classifiers and the support vector machine with linear kernel, and the support vector machine with Radial Basis Function (RBF) kernel. Testing the proposed method on Computer Vision and Biometric Laboratory (CVBL) data indicated that the recognition rate, at best for the illumination challenge in the whole face and a part of the face is 98.6 % and 97.9% respectively. تفاصيل المقالة

تولید سلول‌های اضافی اغلب تشکیل توده‌ای از بافت را می‌دهند که به آن تومور اطلاق می‌شود. تومورها می‌توانند عملکرد صحیح مغز را مختل کنند و حتی منجر به مرگ بیمار گردند. یکی از راه‌های تشخیصی غیرتهاجمی برای این بیماری تصویر‌برداری تشدید مغناطیسی (MRI) می‌باشد. توسعه‌ی یک سیستم تشخیصی اتوماتیک یا نیمه‌اتوماتیک به کمک کامپیوتر در درمان‌های پزشکی مورد نیاز است. الگوریتم‌های متعددی برای تشخیص تومور بکار گرفته شده است که هرکدام دارای مزایا و معایب خاص خودش است. در این پژوهش، از تلفیق روش‌های تقسیم‌بندی سوپرپیکسل و طبقه‌بندی RVM، یک روش اتوماتیک برای پیدا کردن محدوده دقیق ناحیه تومور در تصویر MRI ابداع نموده است. الگوریتم مورد‌استفاده در روش سوپرپیکسل، الگوریتم SLIC است که برای هر سوپرپیکسل 13 ویژگی آماری و شدت روشنائی، محاسبه شده و در نهایت توسط الگوریتم طبقه‌بندی RVM روشی آموزش داده می‌شود که بتواند در هر تصویر MRI مغز، قسمت تومور را از غیر‌تومور تشخیص دهد.در این تحقیق از مجموعه داده BRATS2012 و از تصاویر با وزن FLAIR استفاده شده است و نتایج بدست آمده با نتایج BRATS2012 مقایسه گردیده است و ضرایب هم‌پوشانی Dice، BF score و Jaccard به ترتیب 0.898 ، 0.697 و 0.754 بدست آمده است. تفاصيل المقالة

در مراحل مختلف تولید پارچه، خرابی‌هایی متعددی برسطح پارچه ظاهر می‌شود. با چشم پوشی از دلایل ایجاد خرابی ‌ها، تشخیص دقیق انواع آن‌ها به طبقه بندی صحیح پارچه کمک می‌کند و در نتیجه درصد بالایی از فرآیند کنترل کیفیت را به خود اختصاص می‌دهد. کنترل کیفیت پارچه‌ به ‌منظور بهبود کیفیت محصول و حفظ بازار رقابتی از اهمیت بالایی برخوردار است. همچنین شناسایی نواحی معیوب در روش های خودکار از اهمیت ویژه ای برخورداراست. در این مقاله، یک روش جدید، جهت ناحیه بندی نواحی معیوب پارچه ، مبتنی بر خوشه بندی و همچنین عملگرهای ریخت شناسی ارائه شده است. در روش پیشنهادی، پس از پیش پردازش های لازم جهت بهبود کیفیت تصویر، در مرحله اول روی تصویر خوشه بندی صورت می‌گیرد، تا نواحی مشابه ایجاد شوند. سپس عملگرهای ریخت شناسی به کار گرفته می شوند تا ناحیه معیوب استخراج شود. استفاده از ترکیب هوشمندانه عملگرهای ریخت شناسی، سبب شناسایی دقیق نواحی معیوب درتصویر پارچه شده است. نمایش ناحیه معیوب به کمک الگوریتم کانتور فعال صورت می گیرد. اگرچه تاکنون روش های متعددی همچون الگوهای محلی دودویی و سایر روش ها ارائه شده است، اما سرعت شناسایی این الگوریتم ها پایین بوده و پیچیدگی محاسباتی بالایی دارند. روش پیشنهادی روی پایگاه داده CMUPIE، پیاده سازی شده و به کمک معیارهای صحت و دقت ارزیابی شده است. صحت شناسایی نواحی معیوب در روش پیشنهادی، 82/93درصد و دقت روش پیشنهاد شده، 33/98 درصد حاصل گردیده است که در مقایسه با روش های مشابه، بهبود چشم گیری داشته است تفاصيل المقالة

مدار تمام جمع کننده، به دلیل توانایی در پیاده‌سازی چهار عمل اصلی محاسباتی (جمع، تفریق، ضرب و تقسیم) به عنوان یکی از مهمترین و پرکاربردترین بخش‌های اصلی پردازنده‌های دیجیتالی در طرّاحی مدارهای مجتمع، شناخته می‌شود. بدین منظور، در این مقاله تلاش شده است که سلول تمام جمع‌کننده‌ی جدیدی با بهره‌گیری از تکنولوژی ترانزیستورهای نانولوله‌ی کربنی، جهت دستیابی به مداری با عملکردی مناسب و توان مصرفی کم، ارائه گردد. طرح پیشنهادی از 12 ترانزیستور CNTFET که با استفاده از منطق ترانزیستورهای عبور به هم متصل شده‌اند، تشکیل شده است. ترانزیستورهای نانولوله‌ی کربنی در توان مصرفی و سرعت عملکرد، برتری قابل توجهی نسبت به ترانزیستورهایMOSFET از خود نشان می‌دهند. شبیه‌سازی طرح پیشنهادی، با استفاده از نرم افزار Hspice و بر مبنای مدل CNTFET، با ولتاژ اعمالی V65/0 در سه فرکانس و سه مقدار خازن بار متفاوت، انجام می‌شود و نتایج به دست آمده، برتری طرح پیشنهادی را نسبت به مدارهای نظیر ارائـه شده در مقالات پیشین، اثبـات می‌کند تفاصيل المقالة