Search

In this paper we have developed a Matlab/Simulink based model for monitoring a contact in a video surveillance sequence. For the segmentation process and corect identification of a contact in a surveillance video, we have used the Horn-Schunk optical flow algorithm. The position and the behavior of the correctly detected contact were monitored with the help of the traditional Kalman filter. After that we have compared the results obtained from the optical flow method with the ones obtained from the Kalman filter, and we show the correct functionality of the Kalman filter based tracking. The tests were performed using video data taken with the help of a fix camera. The tested algorithm has shown promising results.

Estimating its ego-motion is one of the most important capabilities for an autonomous mobile platform. Without reliable ego-motion estimation no long-term navigation is possible. Besides odometry, inertial sensors, DGPS, laser range finders and so on, vision based algorithms can contribute a lot of information. In this paper, we want to give an integrated overview of stereo odometry. We want to emphasize the fact that stereo odometry is a chain of several single subprocesses where each relies on its predecessor’s results. The key to a more accurate and efficient stereo odometry lies in an integrated analysis of its single subprocesses and the many algorithms available. [ABSTRACT FROM AUTHOR]

This paper presents an approach to handle with elements from GF(2n), in hardware implementation with minimum costs of area. The method is described by exemplifying with minimum costs of area. The method is described by exemplifying the practical implementation of the BCH(n,k,t) scheme over GF(2n), where n is large (n > 6), on reconfigurable FPGA hardware with minimum costs of area. There are many papers in the open literature which presents hardware implementations of algorithms over GF(2n) but none of them addresses the problem of hardware resources employed. There are many situations in which an area optimized implementation is more suitable than a speed optimized implementation. [ABSTRACT FROM AUTHOR]

The paper introduces an algorithm which improves the value of the real giga floating point operations per second (GFLOPS) for matrix multiplication algorithm on Graphical Process Unit-GPU by overlapping the data transfers between (CPU) and the device (GPU) with the kernel execution. The input matrices are divided into n sections and the output matrix into n² sections. Streams are used to perform simultaneous data transfers and kernel executions in order to hide the memory copy operations. The results show that improved execution times and GFLOP values are obtained. The optimum value of n depends mainly on the matrix dimension and on the GPU type. [ABSTRACT FROM AUTHOR]

The paper compares the single and double floating point precision performance of NVIDIA Tesla GPUs. Double precision is crucial for the accuracy of some applications containing scientific computation. Three representative applications have been chosen in order to compare the performances: matrix multiplication, incompressible Navier-Stokes equations and the steady state heat conduction problem. The expected values of performance decrease lie between two for bandwidth limited applications and eight for compute limited applications. Based on the conducted experiments, the type of each application has been identified and the decreases in performance have confirmed the expected theoretical values. [ABSTRACT FROM AUTHOR]