Applied Mechanics and Materials Vols. 687-691

Abstract: Combining level set image segmentation with the prior shape information, we proposed an improved prior shape model in face recognition. Firstly, we introduced a local tensile invariant to X and Y direction and a shear invariant based on the shape statistics. Then, shape energy term with rotation, scale, shear and translation invariance was reconstructed in level set C-V model. The new model considers global and local image change and makes face contour evolving stably. Experimental results demonstrate that our model can segment obscured face images effectively.

Abstract: This article made a in-depth research of the face detection with the method of integral image, which is based on image capture and recognition technology, and designed the hardware circuit and software program development framework. Designed hardware circuit platform around the Cortex-A8 core processor in hardware, which was exclusively for the camera driver, face recognition and image capture. Prorammed face detection code with QT, and finally transplanted the face detection program to ARM board. Results show that the system has a high identification rate correctly and a good real-time performance under normal lighting conditions after a certain sample size of the test.

Abstract: Aiming at the de-noising of GPR echo signal, a de-noising method based on EEMD and wavelet is presented. First the echo signal data is processed with EEMD and yields IMF components. Then the IMF components which indicate noise are subtracted. Next, the high frequency IMF components of the remaining are subjected to wavelet threshold. Finally, the signal is reconstructed using the de-noising IMF and low frequency IMF to realize signal de-noising. Compared with other commonly used methods, EEMD-wavelet method has improvement on SNR. The experiment results show its effectiveness and feasibility in GPR de-noising.

Abstract: First disease spot color and texture features were extracted from barley field images in Gansu, and the feature vectors were used as input vector to establish barley diseases classifier model. Then the neural network was applied to rain classified model with collected images as training set. Finally, two groups of random selected images as test sets were used to perform classified verification experiments. The experimental results show that the overall accuracy of barley dis-eases recognition model is above 86.7%. Therefore, Barley disease image recognition based on neural net-work provides a new technology for the classified treatment of barley diseases.

Abstract: ECG signal contains abundant information of human heart activity. It is important basis of doctors’ diagnose. With the development of computer technology, computer aided analysis has been widely applied in the field of ECG analysis. Most of the traditional method is based on single classifier and too complex. Also, the accuracy is not high. This paper focuses on ECG heart beat classification, extracting different types of feature, training different classifiers by vector model and support vector machine (SVM), merging the result of multiple classifiers. In this paper, we used the advanced voting method (voting by weight) to fusion the result of different classifier, having compared it with the traditional voting method.It performed better than traditional method in term of accuracy

Abstract: Loudspeaker equalization aims at mitigating the distortions introduced in the process of audio reproduction. Conventional equalization schemes focus mainly on the axial room impulse response, and thus cannot fulfill the requirements of near-filed applications such as desktop loudspeakers. This paper describes a novel equalization scheme, in which the problem is partially solved by taking the binaural perception difference into consideration.

Abstract: In this paper we present a compound anisotropic diffusion filter algorithm to apply edge sensitive ICOV operator in NCD model. According to the correlation coefficient of the ICOV operator, we obtain effective nonlinear denoising. The experiment have validated that our algorithm have better effect in smoothing and better ability in edge preservation.

Abstract: Person re-identification, which means matching person across non-overlapping cameras in a surveillance camera network, has attracted more and more attention. A lot of metric learning based methods, which generally learn a new distance function under two pair-wise constrains, i.e. similar constrain and dissimilar constrain, were proposed to address the challenging problem due to significant appearance variances caused by pose changes, lighting variations and image resolution differences. However, these methods attempt to satisfy all similar constrains and dissimilar constrains, which may be conflict and cannot be simultaneously satisfied in the practical application. In this paper, we propose a new local metric learning method based KISS metric learning. Comparative experiments conducted on three public standard datasets have shown the promising prospect of the proposed method.

Abstract: for the mutual restriction problem of precision and efficiency of current WIFI positioning technology, we propose a locating algorithm combining the Location Fingerprint with the Physical Decay Model and carry out de-noising treatment during the data collection. Use direct physical decay model to position within a tolerable error range. When the error exceeds the threshold, combined with location fingerprint algorithm, we use KNN for further exact match. Experients show that this method can effectively reduce the errors caused by unstable RSSI and improve the positioning accuracy and efficiency.

Abstract: The vibration of the string is a series of different frequency and different amplitude and phase of harmonic vibration superposition. These different frequencies are including fundamental frequency, double or triple frequency, and more integer time’s frequency. It is composed of string segment of vibration. The piano sound is spread by the string vibration, through the string to the soundboard that is by the soundboard to air, and then is introduced to the process of the human ear. String vibration is sound energy generated link, its energy is the key to the size of the sound vibration energy; therefore, the sound propagation attenuation problem in the first part is crucial, how to improve the strings attenuation becomes a key problem to be solved. This article aim is that the application of the string vibration theory in the piano strings sound reconstruction and its implementation are discussed.