Yazar "Baykal, Ibrahim Cem" seçeneğine göre listele

Listeleniyor 1 - 11 / 11
  • [ X ]
    Öğe
    Diftogram: A Better Texture Discriminator
    (IEEE, 2018) Baykal, Ibrahim Cem
    This article aims to design the most efficient texture discriminator by proposing two alterations to the existing algorithms. The first argument presented in this article is that the direct use of Unser's difference histogram (not sum) as input to a classifier yields more successful results than Haralick's method. The second proposal is the introduction of two different methods for the selection of the pixel distance and the orientation. Both mathematical and empirical evidence is presented, proving that these methods improve the texture discrimination rate by %250-%300 while reducing the computational complexity by %50.
  • [ X ]
    Öğe
    Diftogram: A Better Texture Discriminator
    (Institute of Electrical and Electronics Engineers Inc., 2019) Baykal, Ibrahim Cem
    This article aims to design the most efficient texture discriminator by proposing two alterations to the existing algorithms. The first argument presented in this article is that the direct use of Unser's difference histogram (not sum) as input to a classifier yields more successful results than Haralick's method. The second proposal is the introduction of two different methods for the selection of the pixel distance and the orientation. Both mathematical and empirical evidence is presented, proving that these methods improve the texture discrimination rate by %250-%300 while reducing the computational complexity by %50. © 2018 IEEE.
  • [ X ]
    Öğe
    Frequency Domain Improvements to Texture Discrimination Algorithms
    (Science & Information Sai Organization Ltd, 2023) Baykal, Ibrahim Cem
    As the production speeds of factories increase, it becomes more and more challenging to inspect products in real time. The goal of this article is to come up with a computationally efficient texture discrimination algorithm by first testing their ability to localize defects and then increase their efficiency by removing less effective parts of them. Therefore, abilities of the most popular texture classification algorithms such as the GLCM, the LBP and the SDH to localize defects are tested on different datasets. These tests reveal that, on small windows GLCM and SDH perform better. Frequency properties of the textures are used to fine-tune the parameters of these algorithms. Further experiments on three different datasets prove that the accuracy of the algorithms are increased almost twice while decreasing the processing time considerably.
  • [ X ]
    Öğe
    Frequency Domain Improvements to Texture Discrimination Algorithms
    (Science and Information Organization, 2023) Baykal, Ibrahim Cem
    As the production speeds of factories increase, it becomes more and more challenging to inspect products in real time. The goal of this article is to come up with a computationally efficient texture discrimination algorithm by first testing their ability to localize defects and then increase their efficiency by removing less effective parts of them. Therefore, abilities of the most popular texture classification algorithms such as the GLCM, the LBP and the SDH to localize defects are tested on different datasets. These tests reveal that, on small windows GLCM and SDH perform better. Frequency properties of the textures are used to fine-tune the parameters of these algorithms. Further experiments on three different datasets prove that the accuracy of the algorithms are increased almost twice while decreasing the processing time considerably © 2023, International Journal of Advanced Computer Science and Applications.All Rights Reserved.
  • [ X ]
    Öğe
    Improving the Hough Transform through Three New Morphological Operators
    (Institute of Electrical and Electronics Engineers Inc., 2019) Baykal, Ibrahim Cem
    Despite being one of the most famous line detectors, Hough Transform has reliability issues. Hough Transform works better on long lines. This article introduces three new binary morphological operators that stretches binary edge images. The first of these operators thins the output of the binary edge image so that the second one can stretch it. The third operator, smooth, is designed for smoothing the lines that are stretched. Results show that these two operators improve Hough Transform results considerably at a very low computational expense. The author publicly shared the executable of the algorithm and the dataset on the Github to prove the efficiency of the method. © 2019 IEEE.
  • [ X ]
    Öğe
    Inspection of Screw Holes on Machine Parts Using Robot Vision
    (IEEE, 2018) Baykal, Ibrahim Cem
    Machine parts have numerous screw holes on them to enable them to be assembled with other parts. If those holes are not placed precisely at the correct locations, the assembly process cannot be accomplished. Quality control departments spend considerable time and effort trying to inspect machine parts manually. This article describes a robot vision based inspection system that constructs a 3-D model of the machine part, including the screw holes and then compares it to a perfect model. A camera connected to a robot arm takes pictures of the machine part from different angles allowing the computer to extract straight lines to be used as features for 3-D reconstruction.
  • [ X ]
    Öğe
    Inspection of Screw Holes on Machine Parts Using Robot Vision
    (Institute of Electrical and Electronics Engineers Inc., 2019) Baykal, Ibrahim Cem
    Machine parts have numerous screw holes on them to enable them to be assembled with other parts. If those holes are not placed precisely at the correct locations, the assembly process cannot be accomplished. Quality control departments spend considerable time and effort trying to inspect machine parts manually. This article describes a robot vision based inspection system that constructs a 3-D model of the machine part, including the screw holes and then compares it to a perfect model. A camera connected to a robot arm takes pictures of the machine part from different angles allowing the computer to extract straight lines to be used as features for 3-D reconstruction. © 2018 IEEE.
  • [ X ]
    Öğe
    Performance Comparison of Texture Classifiers on Small Windows
    (Institute of Electrical and Electronics Engineers Inc., 2019) Baykal, Ibrahim Cem
    A textured image has both random and periodic components. Texture is defined as the statistical variance of the pixel values with respect to each other therefore, it is impossible to classify a single pixel of any texture. They must be considered as a group or simply as a window. There has to be enough pixels in that window to define the properties of that texture. The smaller the window is, the harder for the texture classifier to recognize it. On the other hand, using smaller windows reduce the processing power requirements and make it easier to localize defects or segment textures. Consequently, the aim of this article is to evaluate the performance of popular texture classifiers on different window sizes. In this article, the performance of a Support Vector Machine classifier versus an Artificial Neural Network is compared as well. Results show that GLCM-16 combined with an Artificial Neural Network is superior to GLCM-256 and LBP variants on small windows. © 2019 IEEE.
  • [ X ]
    Öğe
    Real-Time Inspection of MDF Fiber Spread Uniformity
    (IEEE, 2018) Baykal, Ibrahim Cem; Yeltekin, A. T.; Budak, O.; Turan, E.
    Wood pieces are turned into very thin fibers for the production of the Medium Density Fiberboard (MDF). These fibers are than spread on a conveyor belt before going into the hot press. These fibers must be distributed uniformly on the conveyor belt in order to produce high quality boards. When the mechanical spreader clogs up, the fiber is spread unevenly, creating trench like vertical craters. This article describes the methods and the algorithms to detect those craters in real-time. The method consists of special lighting to create shadows of those craters and algorithms to detect the shadows and the change in the texture. This article introduces a new method; direct input of Unser's difference histograms to Neural Networks to detect the patterns caused by the uneven spread. The article then shows how existing feature elimination methods can be modified specific to this method to achieve %35 reduction in processing power compared to the conventional methods.
  • [ X ]
    Öğe
    Real-Time Inspection of MDF Fiber Spread Uniformity
    (Institute of Electrical and Electronics Engineers Inc., 2019) Baykal, Ibrahim Cem; Yeltekin, A.T.; Budak, O.; Turan, E.
    Wood pieces are turned into very thin fibers for the production of the Medium Density Fiberboard (MDF). These fibers are than spread on a conveyor belt before going into the hot press. These fibers must be distributed uniformly on the conveyor belt in order to produce high quality boards. When the mechanical spreader clogs up, the fiber is spread unevenly, creating trench like vertical craters. This article describes the methods and the algorithms to detect those craters in real-time. The method consists of special lighting to create shadows of those craters and algorithms to detect the shadows and the change in the texture. This article introduces a new method; direct input of Unser's difference histograms to Neural Networks to detect the patterns caused by the uneven spread. The article then shows how existing feature elimination methods can be modified specific to this method to achieve %35 reduction in processing power compared to the conventional methods. © 2018 IEEE.
  • [ X ]
    Öğe
    Ultrafast line detector
    (SPIE, 2022) Yilmaz, Ismail Can; Baykal, Ibrahim Cem
    This article introduces the fastest line detector so far published in the literature. This algorithm is 14.88 times faster than the next fastest line detector, the EDLines, and 94.7 times faster than the line segment detector. It is hundreds of times faster and more reliable than the Standard Hough Transform. The article also introduces an ingeniously simple method for detecting and combining patterns. The algorithm takes advantage of several look-up tables to recognize and fit straight-line patterns. As the first step, it recognizes any possible 4 × 4 pixel line patterns among the binary edge pixels and then uses several small look-up tables to decide whether the connected patterns form a line or not. It is specially designed for real-time processing of the high-resolution images such as the ones used in computer vision applications. © 2022 SPIE and IS&T.