Search

The segmentation method of classical random walker needs to select seed points artificially, and may cause incomplete or inaccurate segmentation. In view of these problems, this paper presents an image positioning method of automatic random walker based on IFS (Intuitionistic Fuzzy Set). IFS edge detection method is used to get the edge information of image, and then the connected domains are found out in image edge using the morphology method. Select the central pixels of each connected domain as seed points for automatic random walker, and then segment the IFS edge image by random walker. Experimental results show that the proposed method overcomes the application limitation of semi-automatic random walker algorithm, and improves the accuracy of image positioning.

A knowledge reasoning based on-line infrared image monitoring system of the power equipments in substation is designed. The function, feature and structure of the system are introduced. Compare with the usual on-line infrared imaging monitoring system, a knowledge base is added and a knowledge reasoning method is used to diagnose the fault mode of operating equipments. The representation method of knowledge and processing of knowledge reasoning engine are introduced. Our research demonstrates that fault can be diagnosed correctly.

The Smart Home is an important realization of the environments of intelligent electricity. As an essential part of the system, smart home smart sockets and the gateway work together to achieve the function of the on-off control, intelligent control, data acquisition and overload protection, etc. In this paper,the smart socker was designed and composed with the CC430 series processors introduced by TI , the power module, metering module and relay control module .At last, each module was tested accordingly.

In order to eliminate the influences of illumination and face-poses on eye localization, a feasible method is proposed based on the skin color feature and Otsu algorithm. Firstly, we detect the skin color in YCb'Cr' color space. Skin color segmentation principle is used to narrow the search region in human eye detection. Then we convert the segmented image to a binary image by Otsu algorithm and extract the eye region. Finally, the left and right eyes are positioned in the facial area with the binary integral projection. An analysis of the detections reveals that this algorithm has good robustness against changes of illumination and face-pose.

Digital image forgery is a growing problem as the image could be easily manipulated. Digital image forgery detection has recently received significant attention. A fast and efficient blind detection algorithm is presented for the detection of eclosion forgeries. The second generation of Curvelet transform is firstly used to enhance the image, and then Canny operators is used to detect the edge of the enhanced image, finally the eclosion forgeries can be identified from the edge image. Experimental results show that the algorithm has less operation time and can distinguish the weak edges from the eclosion edges effectively. So it has a higher detection precision and efficiency.

Echo hiding is one of the information hiding technologies that uses audio as the carrier. In this paper, a time-spread echo hiding steganalysis algorithm based on ensemble learning is proposed. By analyzing the principle of the time-spread echo hiding and ensemble learning, the steganalytic system of time-spread echo hiding based on AdaBoost ensemble learning algorithm is built. The features that contained cepstrum, the first-order derivatives of cepstrum, second-order derivatives of cepstrum, the mel-cepstrum, combined histogram and moments of high-order statistics are simulated by using SVM and AdaBoost ensemble learning algorithm. Experiment results show that the classification effects of the ensemble learning are better than the SVM.

A new approach was proposed to recognize personal name and location name based on conditional random fields in Chinese language domain. Some interesting features have been proposed, the new probabilistic feature is proposed, which are used instead of binary feature functions, however, it is one of the several differences between this model and the most of the previous CRFs-based model. We also explore several new features in our model, which includes local features, global features, hybrid feature, related features etc; Finally, the linguistics model with multi-knowledge is built. The proposed approach is evaluated with open test method using People's Daily (January, 1998), the evaluation results show that the F-measure is 85.91% for personal name and the F-measure is 93.14% for location name, which prove our approach has a good performance.

Steganalysis is taken as a countermeasure to steganography and is detecting and decoding hidden data within a given media. There has been quite some effort in audio steganalysis for additive embedding model. However, when they distinguish the cover-audio signal with multiplicative noise and the stego-audio signal, results are disappointing. In this paper, a wavelet domain audio steganalysis method for multiplicative embedding model is proposed. The test audio signal is firstly calculated its absolute value and logarithm. Multiplicative noise is changed to additive noise. Then features are extracted. At last, support vector machine (SVM) is utilized as a classifier to distinguish the cover-audio signal and the stego-audio signal. Simulation results show that the detection rates are greater than 94% and the method is effective.

Traditional speech information hiding methods have several disadvantages, for example, constant embedding amplitude, lower speech quality, higher bit error rate. A novel speech information hiding method based on Itakura-Saito measure and psychoacoustic model is proposed. The embedding amplitude can be controlled by Itakura-Saito measure and psychoacoustic model together. The host speech is decomposed by wavelet packet transformation and then mapped into the critical bands. According to the audio masking threshold, the embedding amplitude in each subband can be determined. And then, the adjustment factors can be calculated by Itakura-Saito measure to control the embedding amplitude in each frame so that the speech quality is good. The embedding amplitude can be determined automatically. Experimental results show that the performance of this method is better than that of the traditional methods.

A wavelet domain audio steganalysis method based on principal component analysis (PCA) is proposed. The audio signal is firstly decomposed by 4-level discrete wavelet transform, then the 36 statistical moments of the histogram and the frequency domain histogram for both the audio signal and its wavelet subbands are calculated as features, then the preprocessing of PCA is used on the statistics features and radial basis function (RBF) network is utilized as a classifier. The proposed scheme not only reduces the dimension of the feature vector effectively and simplifies the design of the classifier, but also keeps the detection performance. Then this scheme is utilized to detect the stego-audio signals embedded by wavelet domain LSB, quantization index method (QIM) and addition method (AM). Simulation results show that the performance of our scheme is better than that of the scheme proposed by Xuemin Ru and the detection rates are all greater than 92%.