Melih Cevdet Ince

Silhouette Orientation Volumes for Efficient Fall Detection in Depth Videos

A novel method to detect human falls in depth videos is presented in this paper. A fast and robust shape sequence descriptor, namely the Silhouette Orientation Volume (SOV), is used to represent actions and classify falls. The SOV descriptor provides high classification accuracy even with a combination of simple associated models, such as Bag-of-Words and the Naïve Bayes classifier. Experiments on the public SDU-Fall dataset show that this new approach achieves up to 91.89% fall detection accuracy with a single-view depth camera. The classification rate is about 5% higher than the results reported in the literature. An overall accuracy of 89.63% was obtained for the six-class action recognition, which is about 25% higher than the state of the art. Moreover, a perfect silhouette-based action recognition rate of 100% is achieved on the Weizmann action dataset.

The FIR filter design by using window parameters calculated with GA

In this study, digital filter design was realized helping window functions that their parameter values were calculated using GA. First in the study, the Kaiser window function parameters used many application was found with GA. This values obtained were used in FIR filter design and filter coefficient values which provide desired conditions were calculated. The amplitude responses obtained from available commands and filter amplitude response according to FIR filter coefficient values founded with programme were compared and plotted. The results have shown that the improved programme has been successful. Besides, this study is to provide decreasing of process steps for calculation of window function coefficients used at different filter conditions.

OMP-ELM: Orthogonal Matching Pursuit-Based Extreme Learning Machine for Regression

Abstract Extreme learning machine (ELM) is a recent scheme for single hidden layer feed forward networks (SLFNs). It has attracted much interest in the machine intelligence and pattern recognition fields with numerous real-world applications. The ELM structure has several advantages, such as its adaptability to various problems with a rapid learning rate and low computational cost. However, it has shortcomings in the following aspects. First, it suffers from the irrelevant variables in the input data set. Second, choosing the optimal number of neurons in the hidden layer is not well defined. In case the hidden nodes are greater than the training data, the ELM may encounter the singularity problem, and its solution may become unstable. To overcome these limitations, several methods have been proposed within the regularization framework. In this article, we considered a greedy method for sparse approximation of the output weight vector of the ELM network. More specifically, the orthogonal matching pursuit (OMP) algorithm is embedded to the ELM. This new technique is named OMP-ELM. OMP-ELM has several advantages over regularized ELM methods, such as lower complexity and immunity to the singularity problem. Experimental works on nine commonly used regression problems indicate that the investigated OMP-ELM method confirms these advantages. Moreover, OMP-ELM is compared with the ELM method, the regularized ELM scheme, and artificial neural networks.

Limiting GPR in a two-layer soil model via genetic algorithms

Abstract In this paper, an optimum grounding grid that provides the conditions of GPR E touch and minimum cost in the structures of two-layer soil model is designed and the length of total conductor and the quantity of ground rod are calculated via Genetic Algorithms (GA). A new approach is presented for the calculation of total conductor length. At the same time, the subject regarding in which layer the ground conductors and rods that form the grounding grid in a substation are to be placed in two-layer soil is analysed using GA. With this as the goal, the depth of optimum grid burial is determined. Our study is compared with the design study for a two-layer soil model in the literature. As a result, the high performance of optimum grid design that is achieved using GA is emphasized by varied applications.

Iterative hard thresholding based Extreme Learning Machine

Extreme Learning Machines (ELM) is a new learning algorithm for Single hidden Layer Feed-forward Networks (SLFNs). The ELM has better generalization, rapid training and lower complexity, however, the method suffer from singularity problem and obtaining optimum number of neurons in the hidden layer. In this paper, we considered an IHT for sparse approximation of the output weights vector of the ELM network. The performance evaluation of the proposed method which is called IHT-ELM, was chosen out on four commonly used medical dataset for prediction purposes. The results showed that IHT-ELM has several advantages against the original ELM methods such as obtaining optimum number of neurons and low complexity.

Fall detection with depth-videos

Elderly people, who are living alone are great risk if a fall event occurred. Therefore automatic fall detection systems are in demand for elderly people. In this paper, a depth based fall detection system is proposed. The proposed method consists of shape based fall characterization and a Support Vector Machine (SVM) classifier. Shape based fall characterization is formed with Curvature Scale Space (CSS) features and Fisher Vector (FV) encoding. According to the results obtained from experimental study, the proposed method has better performance than other methods that were published in literature.

The effect of uniform scattered Genetic Algorithms to window function performance

In this study, the effects of Genetic Algorithms (GA) having uniform scattered initial population was analysed to window function performance used at design of finite impulse response filter prefered for many areas. At work, the comparison with amplitude response of Kaiser window having many features was done. With this work, GAs having different initial population were used to obtain more useful window function at calculation of suitable spectral parameter values. The obtained results show that to get result, uniform scattered GA is more successful than having random produced population. Addition, obtained results were shown with figures and were compared with Kaiser window results and finally, it was observed that the improved method for some applications has more useful structure.