Muhammed Fatih Talu

Comparison of Extended-Kalman- and Particle-Filter-Based Sensorless Speed Control

State estimation process is one of the major concerns for controlling and monitoring systems in industry which requires high-cost measurements or unmeasurable variables of nonlinear systems. These drawbacks can be highly eliminated by designing systems without using any kind of sensors. In this paper, sensorless speed control of a dc motor was performed by using extended Kalman filter (EKF) and particle filter (PF). The speed information is estimated by using armature current data measured from a dc motor which is controlled in various speed references with a closed-loop controller. Furthermore, a performance comparison of the EKF and the PF by taking into consideration their estimation errors under the same conditions was realized in a simulation environment. The comparison results showed that the estimation performance of the PF is more accurate but slower than the EKF. The quantitative values of accurateness and slowness are depended on the particle number of the PF. The obtained computation times of the PF having ten particles and the EKF are 180 and 15 μs, respectively.

Continuous rotation invariant features for gradient-based texture classification

Four highly discriminative and continuous rotation invariant methods are proposed.We use the Hessian matrix and Gaussian derivative filters.Verified on the CUReT, KTH-TIPS, KTH-TIPS2-a, UIUC and Brodatz texture datasets. Extracting rotation invariant features is a valuable technique for the effective classification of rotation invariant texture. The Histograms of Oriented Gradients (HOG) algorithm has been proved to be theoretically simple, and has been applied in many areas. Also, the co-occurrence HOG (CoHOG) algorithm provides a unified description including both statistical and differential properties of a texture patch. However, HOG and CoHOG have some shortcomings: they discard some important texture information and are not invariant to rotation. In this paper, based on the original HOG and CoHOG algorithms, four novel feature extraction methods are proposed. The first method uses Gaussian derivative filters named GDF-HOG. The second and the third methods use eigenvalues of the Hessian matrix named Eig(Hess)-HOG and Eig(Hess)-CoHOG, respectively. The fourth method exploits the Gaussian and means curvatures to calculate curvatures of the image surface named GM-CoHOG. We have empirically shown that the proposed novel extended HOG and CoHOG methods provide useful information for rotation invariance. The classification results are compared with original HOG and CoHOG algorithms methods on the CUReT, KTH-TIPS, KTH-TIPS2-a and UIUC datasets show that proposed four methods achieve best classification result on all datasets. In addition, we make a comparison with several well-known descriptors. The experiments of rotation invariant analysis are carried out on the Brodatz dataset, and promising results are obtained from those experiments.

Principal curvatures based rotation invariant algorithms for efficient texture classification

The histograms of oriented gradients (HOG) and co-occurrence HOG (CoHOG) algorithms are simple and intuitive descriptors. However, the HOG and CoHOG algorithms based on gradient computation still have some shortcomings: they ignore meaningful textural properties and are unstable to noise. In this paper, two new efficient HOG and CoHOG methods are proposed. The proposed algorithms are based on the Gaussian derivative filters, and the feature vectors are obtained by means of principal curvatures. The feature vectors are rotation invariant by means of the rotation invariance characteristic of principal curvatures (i.e. eigenvalues). The experimental results on the CUReT, KTH-TIPS, KTH-TIPS2-a, UIUC, Brodatz album, Kylberg and Xu datasets confirm that the developed algorithms have higher classification rates than state-of-the-art texture classification methods. The classification results also demonstrate that the developed algorithms are more stable to noise and rotation than the original HOG and CoHOG algorithms. Two continuous rotation invariant descriptors are proposed.The proposed descriptors are based on principal curvatures which are rotation invariant.The experimental results show the power of the methods particularly in extremely noisy conditions.Our approaches give high classification performance on seven texture databases.

A novel active contour model for medical images via the Hessian matrix and eigenvalues

Abstract This paper presents a new level set formulation for active contour models (ACM). We propose the idea of integrating the eigenvalue information of Hessian matrix into the level set function. By this new level set function, the principal curvature information of images is used to enhance the ability of segmenting boundary regions. The advantages of our model are as follows: firstly, the interior and exterior object boundaries can be segmented with the initial contour being anywhere in the input image. Secondly, this method can work with heterogeneous images. Thirdly, the proposed model can produce smooth and right boundaries of objects having vital importance in medical operations. Extensive experiments demonstrate that the proposed model can obtain better segmentation results.

Real time fabric defect detection system on Matlab and C++/Opencv platforms

In industrial fabric productions, real time systems are needed to detect the fabric defects. This paper presents a real time defect detection approach which compares the time performances of Matlab and C++ programming languages. In the proposed method, important texture features of the fabric images are extracted using CoHOG method. Artificial neural network is used to classify the fabric defects. The developed method has been applied to detect the knitting fabric defects on a circular knitting machine. An overall defect detection success rate of 93% is achieved for the Matlab and C++ applications. To give an idea to the researches in defect detection area, real time operation speeds of Matlab and C++ codes have been examined. Especially, the number of images that can be processed in one second has been determined. While the Matlab based coding can process 3 images in 1 second, C++/Opencv based coding can process 55 images in 1 second. Previous works have rarely included the practical comparative evaluations of software environments. Therefore, we believe that the results of our industrial experiments will be a valuable resource for future works in this area.

Interactive segmentatition implementation

This paper includes a geodesic distance based interactive segmentation algorithms Matlab implementation. In the Matlap implementation, a graphic interface by which user creates foreground and background scribbles over the image has been designed. The first step of the algorithm is based on the modelling of the color values on the scribbles and the calculation of foreground/background probabilities of all pixels in the image. In the second step, the calculated probabilities values are accepted as weight values and the segmentation process has been implemented more precision by using geodesic distance method. It can be show that to able to produce precise segmentation results in real-time, the algorithm can be used especially in medical image segmentation applications.