Fatih Kurugollu

Color image segmentation using histogram multithresholding and fusion

Abstract A novel method for multiband image segmentation has been proposed. The method is based on segmentation of subsets of bands using multithresholding followed by the fusion of the resulting segmentation “channels”. For color images the band subsets are chosen as the RB, RG and BG pairs, whose two-dimensional histograms are processed via a peak-picking algorithm to effect multithresholding. The segmentation maps are first fused by running a label concordance algorithm and then smoothed by a spatial–chromatic majority filter. It is shown that for multiband images, multithresholding subsets of bands followed by a fusion stage results in improved performance and running time.

Image Segmentation by Relaxation Using Constraint Satisfaction Neural Network

Abstract The problem of image segmentation using constraint satisfaction neural networks (CSNN) has been considered. Several variations of the CSNN theme have been advanced to improve its performance or to explore new structures. These new segmentation algorithms are based on interplay of additional constraints, of varying the organization of the network or modifying the relaxation scheme. The proposed schemes are tested comparatively on a bank of test images as well as real world images.

Real-time detection of buried objects by using GPR

In this work the detection process of buried objects is presented utilizing Ground Penetrating Radar (GPR). Background removal algorithm is used to obtain the target signature and correlation process is performed to reveal the reflected target energy Then, a detection warning signal is created depending on a special process. In this work, pulsed GPR system with 1 GHz bandwith is used. Scanning speed is 0.33cm/sec in the sweeping direction and this process is repeated in the walking direction with 4 cm spatial resolution.

A comparison of visual target tracking methods in noisy environments

Motion detection and tracking are useful in computer vision. This work represents a comparison of several methods for target tracking in noisy environments using motion-energy and template location based approaches. In order to remove the registration noise, a morphological filter has been adopted. Since the scale and rotation of the target change, a template update method has been used. The performance of different methods to compare their respective speed and accuracy have been evaluated and the results obtained are presented.

MAP segmentation of color images using constraint satisfaction neural network

An improved segmentation algorithm is proposed, which implements the MAP estimation of the label field using a Constraint Satisfaction Neural Network (CSNN). It uses the advantages of stochastic relaxation with those of Gauss-Markov Random Field (GMRF) models. The performance of the algorithm is compared vis-a-vis alternate relaxation schemes using both synthetic and real images.

Quality inspection in PCBs and SMDs using computer vision techniques

This paper addresses the task of automating the visual inspection of printed circuit board with through hole technique, and surface mount devices (SMD). We have focused on the following quality control of the mounted component using either technology: component twist, or pin defects. A comparison of binarization techniques as well as twist angle estimation methods have been carried out. For inspection of pin defects various morphological image processing techniques have been applied.<<ETX>>

Advanced educational parallel DSP system based on TMS320C25 processors

Abstract This paper describes the design, application, and evaluation of a user friendly, flexible, scalable and inexpensive Advanced Educational Parallel (AdEPar) digital signal processing (DSP) system based on TMS320C25 digital processors to implement DSP algorithms. This system will be used in the DSP laboratory by graduate students to work on advanced topics such as developing parallel DSP algorithms. The graduating senior students who have gained some experience in DSP can also use the system. The DSP laboratory has proved to be a useful tool in the hands of the instructor to teach the mathematically oriented topics of DSP that are often difficult for students to grasp. The DSP laboratory with assigned projects has greatly improved the ability of the students to understand such complex topics as the fast Fourier transform algorithm, linear and circular convolution, the theory and design of infinite impulse response (IIR) and finite impulse response (FIR) filters. The user friendly PC software support of the AdEPar system makes it easy to develop DSP programs for students. This paper gives the architecture of the AdEPar DSP system. The communication between processors and the PC-DSP processor communication are explained. The parallel debugger kernels and the restrictions of the system are described. The programming in the AdEPar is explained, and two benchmarks (parallel FFT and DES) are presented to show the system performance.

Image Segmentation Based on Boundary Constraint Neural Network

Publisher Summary Artificial neural networks based image segmentation methods have gained more acceptance over other methods because of their distributed architectures allowing real-time implementation. Another important advantage of the neural networks is that their robustness structures to the unexpected behavior of input image such as noise. On the other hand, the disadvantage of the neural networks is that the learning phase could be too long, and the resulting segmentation has a noisy boundary. The chapter also investigates a neural network based image segmentation method called constraint satisfaction neural network. It proposes a modification of constraint satisfaction neural network (CSNN) to alleviate both problems. It has been observed that when the edge field is brought in as a constraint, the convergence improves and the boundary noise is reduced.