فهرست مطالب

Computer and Robotics - Volume:14 Issue: 2, Summer and Autumn 2021

Journal of Computer and Robotics
Volume:14 Issue: 2, Summer and Autumn 2021

  • تاریخ انتشار: 1400/11/28
  • تعداد عناوین: 6
|
  • Mahin Tasnimi, HamidReza Ghaffari * Pages 1-9

    Diagnosing benign and malignant glands in thyroid ultrasound images is considered as a challenging issue. Recently, deep learning techniques have significantly resulted in extracting features from medical images and classifying them. Convolutional networks ignore the hierarchical structure of entities within images and do not pay attention to spatial information as well as the need for a large number of training samples. Capsule networks consist of different hierarchical capsules equivalent to the same layers in the CNN neural network. This study tried to extract textural features using a deep learning model based on a capsule network. Thyroid ultrasound images were given to the capsule network as input data, and finally the features learned in the capsule network were used to teach the Support Vector Machine classifier, in order to diagnose thyroid cancer. Experimental results showed that the proposed method with 98% accuracy has achieved better results compared to convolutional networks.

    Keywords: Thyroid Gland, convolutional neural network (CNN), deep learning, Feature Extract, Capsule Network
  • Ehsan Amiri, Ahmad Mosallanejad *, Amiri Sheikhahmadi Pages 11-19
    Copy-Move is a technique widely used in digital image tampering, meaning Copy Move Forgery Detection (CMFD) is still significant research. This paper proposes an optimal keypoint in SIFT (OKSIFT). The OKSIFT method produces images of different sizes and different sigma’s. Then with the help of the Gaussian difference (DoG) method, the maximum and minimum keypoints are calculated. When selecting the optimal keypoints, the absolute value of the second sentence will be used instead of using the Taylor expansion binomial series. First, the keypoints lose their dependence on the blurred regions, and secondly, more keypoints appear at the main edges. In the localization process of the region, considering the cases of multiple copies, method g2NN has been used to compare the keypoints. This method reduces the complexity of keypoint calculations and gives a better answer. Experimental results based on precision, recall, and F1 criteria show that the proposed method, with good robustness, works better than some advanced methods.
    Keywords: Copy-Move Forgery Detection, SIFT, new optimal keypoint in SIFT, Gaussian filter
  • Mojtaba Jahanian *, Abbas Karimi, Faraneh Zarafshan Pages 21-27
    Clustering is one of the essential machine learning algorithms. Data is not labeled in clustering. The most fundamental challenge in clustering algorithms is to choose the correct number of clusters at the beginning of the algorithm. The proper performance of the clustering algorithm depends on selecting the appropriate number of clusters and selecting the optimal right centers. The quality and an optimal number of clusters are essential in algorithm analysis. This article has tried to distinguish our work from other writings by carefully analyzing and comparing existing algorithms and a clear and accurate understanding of all aspects. Also, by comparing other methods using three criteria, the minimum internal distance between points of a cluster and the maximum external distance between clusters and the location of a cluster, we have presented an intelligent method for selecting the optimal number of clusters. In this method, clusters with the lowest error and the lowest internal variance are chosen based on the results obtained from the research.
    Keywords: clustering algorithms, K-means, Clustering, the optimal number of clusters
  • Razieh Asgarnezhad *, Karrar Ali Mohsin Alhameedawi Pages 29-40
    Chronic Kidney Disease is one of the most common metabolic diseases. The challenge in this area is a pre-processing problem. Artificial Intelligence techniques have been implemented over medical disease diagnoses successfully. Classification systems aim clinicians to predict the risk factors that cause Chronic Kidney Disease. To address this challenge, we introduce an effective model to investigate the role of pre-processing and machine learning techniques for classification problems in the diagnosis of Chronic Kidney Disease. The model has four stages including, Pre-processing, Feature Selection, Classification, and Performance. Missing values and outliers are two problems that are addressed in the pre-processing stage. Many classifiers are used for classification. Two tools are conducted to reveal model performance for the diagnosis of Chronic Kidney Disease. The results confirmed the superiority of the proposed model over its counterparts.
    Keywords: Pre-processing, chronic kidney disease, Classification, Machine Learning Techniques
  • Sahar Azizi, Mohammad Menhaj *, Mohammad Norouzi Pages 41-52
    Quadrotor is one of the types of flying robots that has attracted the attention of researchers due to its simple structure and perpendicular flight capability. This paper presents a new method based on machine vision for correct window detection, in smoothly unknown environments. One of the challenges of controlling the Quadrotor path in unknown environments is actually accurate window identification for passing through it. In this study, quadrotor Parrot Bebop2 is used which is equipped with a camera. Also, an algorithm is proposed to perform image processing to identify the window in the environment and control the quadrotor's trajectory, which is implemented on the quadrotor. This method consists of three parts: preprocessing, diagnosis and identification. First, by applying image processing algorithms, we improve the image and delete the data unrelated to the target, and then we use a smart machine vision algorithm to extract information. Furthermore, to control the quadrotor route, a proportional-integral-derivative controller is designed and implemented using Ziegler and Nichols method, which will take place during a real indoor flight in an automated tracking. According to the obtained results, it can be concluded that the use of flying robots can have positive results in military processes and assistance to people in a short time.
    Keywords: Machine Vision, image processing, window detection, Zeigler, Nichols, Quadrotor
  • Mahyar Abbaszadeh, Rezvan Abbasi * Pages 53-66
    The increase in the power generated by the wind hashad effects on the performance of the power system incases such as power quality, safety, stability, andvoltage control. The wind turbines are used to generateelectrical energy from wind. They can work in fixedor variable speeds. The asynchronous generator isdirectly connected to the grid for the fixed-speed windturbines. In order to connect the DFIG (Doubly-FedInduction Generator) to the grid, this machine must beable to integrate its generated power into the grid in aspecific voltage (the grid voltage level). The mainDFIG controlling method is the use of field-orientedvector control for regulating the rotor flux. The DFIGvector control consists of two main parts as grid sideconverter control and rotor side converter control. Therotor side converter is used to control the grid outputpower. This converter regulates the power factors inthe terminals, and actually restores the generatedpower deviation from the reference power through thePID controllers, besides guaranteeing the stability ofthe induction generator. In the current study, the powerwas controlled through the determination of the PIDoptimal coefficient of the rotor and grid sidescontrollers and the gray wolf algorithm in theMATLAB software. In addition, the stability of thesmall signal of the grid equipped with the doubly-fedwind generator in the wind speed turbulenceconditions was optimized to satisfy the requiredcriteria in output active and reactive power of a DFIG.From the simulation results it is observed that theproposed controller yields better results whencompared to other methods in literature in terms ofperformance index.
    Keywords: Optimization, Doubly-Fed Induction, Generator, Wind turbine, gray wolf algorithm