2015, Vol.20, No.1

Journal of The Korea Society of Computer and Information 2023 KCI Impact Factor : 0.65

pISSN : 1598-849X / eISSN : 2383-9945
https://journal.kci.go.kr/jksci

In this paper, we propose an efficient parallel implementation method of a widely used complexfour-stage fire detection algorithm using a graphics processing unit (GPU) to improve the performance ofthe algorithm and analyze the performance of the parallel implementation method. In addition, we use seven different resolution videos (QVGA, VGA, SVGA, XGA, SXGA+, UXGA, QXGA) as inputs of thefour-stage fire detection algorithm. Moreover, we compare the performance of the GPU-based approachwith that of the CPU implementation for each different resolution video. Experimental results using fivedifferent fire videos with seven different resolutions indicate that the execution time of the proposed GPUimplementation outperforms that of the CPU implementation in terms of execution time and takes a25.11ms per frame for the UXGA resolution video, satisfying real-time processing (30 frames per second,30fps) of the fire detection algorithm.

Typically welded bellows are checked with human eye and microscope, and then go through leakagetest of gas. The proposed system alternates these heuristic techniques using support vector machines. Image procedures in the proposed method can cover the irregularity problem induced from human being. To get easy observation through microscope, 3D display system is also exploited. Experimental resultsfrom this automatic measurement show the welding detection is done within one tenth of permitted errorrange.

Feature normalization as a pre-processing step has been widely used to reduce the effect of differentscale in each feature dimension and error rate in classification. Most of the existing normalization methods,however, do not use the class labels of data points and, as a result, do not guarantee the optimality ofnormalization in classification aspect. A supervised rank normalization method, combination of ranknormalization and supervised learning technique, was proposed and demonstrated better result than others. In this paper, another technique, training sample selection, is introduced in supervised feature normalization to reduce classification error more. Training sample selection is a common technique forincreasing classification accuracy by removing noisy samples and can be applied in supervisednormalization method. Two sample selection measures based on the classes of neighboring samples andthe distance to neighboring samples were proposed and both of them showed better results than previoussupervised rank normalization method.