Search

The data sparsity and prediction quality are recognized as the key challenges in the existing recommender Systems. Most of the existing recommender systems depend on collaborating flitering (CF) method which mainly leverages the user-item rating matrix representing the relationship between users and items. However, the CF-based method sometimes fails to provide accurate information for predicting recommendations as there is an assumption that the relationship between attributes of items is independent and identically distributed. In real applications, there are often several kinds of coupling relationships or connections existed among users or items. In this paper, we incorporate the coupling relationship analysis to capture the under-discovered relationships between items and aim to make the ratings more reasonable. Next, we propose a neighborhood-based matrix factorization model, which considers both the explicit and implicit correlations between items, to suggest the more reasonable items to user. The experimental evaluations demonstrate that the proposed algorithms outperform the state-of-the-art algorithms in the warm- and cold-start settings.

A new face image feature extraction and recognition algorithm based on Scale Invariant Feature Transform (SIFT) and Local Linary Patterns (LBP) is proposed in this paper. Firstly, a set of keypoints are extracted from images by using the SIFT algorithm; Secondly, each keypoint is described by LBP patterns; Finally, a combination of the global and local similarity is adopted to calculate the matching results for face images. Calculation results show that the algorithm can reduce the matching dimension of feature points, improve the recognition rate and perspective; it has nice robustness against the interferences such as rotation, lighting and expression.

In this paper it proposes a meaningful digital watermarking algorithm for remote sensing image based on DFT and watermarking segmentation. At First, it normalizes the host remote sensing image and determines an invariant centroid, then selects a square area around the invariant centroid for watermark embedding. Next, it generates a pseudo-random sequence as digital watermarking, and divides it into two parts. Finally it applies DFT to the selected region and embeds watermark into DFT phase and amplitude components of the selected square area of the host remote sensing image. Experiments have shown the algorithm’s characteristics of good robustness, simple calculation, easy realization, and extracting watermark without the original remote sensing image, and it has value of copyright protection for remote sensing images.

Edge is one of the most important image features, and it is the base of follow-up measurement and shape registration. In order to get better edge information, an edge detection method of block distance is proposed in this paper. Firstly, by calculating distance between the pixels within sub-block regions, the graph of Gaussian weighted distance is obtained. Comparing with the original figure, it not only can better highlight the edge contour, but also can be uniformed background gray. Secondly, in order to reduce the complexity of the algorithm, edge detection method of block distance combining with summed area table is raised. Finally, implementation process of this algorithm is raised. Experimental results show that this method can achieve high speed of edge detection with higher accuracy in extracting edge and corner information of image.

In order to extract characteristics of face by making full use of LBP and improve its "adaptive ability", we proposed an algorithm based on global and local fusion LBP. First, we will extract overall face feature histogram with LBP, then segment the image into blocks, extract each LBP histogram feature, then combine the global and local features according to certain order , and we regard it as the total character of image. Finally, we use the LS- SVM (least squares support vector machine) identifying and training samples of face image to improve the speed of recognition the experiment on ORL face database shows that the algorithm has high recognition rate and has improved recognition rate of the face.

This paper presents a point cloud reconstruction algorithm which based on SVR(support vector regression) . Firstly, the point cloud data pre-processing, filter out noise points. Then train the point by SVR , and we can get the function of surface expression. Finally, using the Marching Cube algorithm to visualize the implicit function. Experimental results show that the algorithm is more robust and more efficient.

Computer simulation of the flame is difficult to achieve real-time and realistic problem, proposing a fire simulation method based on fluid model and GPU general computing combining. The method is based on the incompressible flame, low-density, non-sticky and so on. Semi-Lagrange method is using to solve the fluid equations, using volume rendering based on 3D texture to rendering the flame. Then, using the method of energy-spectrum of radiation to control the color of the flame, and using the GPU to accelerate in parallel, balance realistic and real time.

Skin color detection is an important in computer vision.This work presents a new efficient method for skin color detection based on an improved direct least square ellipse fitting in the space CrCbCg. The color distribution statistics of three-dimensional CrCbCg is obtained by fitting the color distribution ellipse boundary in the Cr-Cb, Cr-Cg, Cb-Cg plane exactly ,finally ,the criteria based on the statistics is used to detect skin color area accurately. Experimental results show that this method improves the robustness of the skin color detection to complex environment.

With the processing of CMOS technology, the technology of video acquisition based on CMOS is becoming a new trend. However, manyCMOS camera chip is not supported by the newest Linux kernel yet. In this paper, the method of designing the CMOS camera driver based on S3C2440 developing board with the embedded Linux environment is introduced. SCCB is a distinguishing feature of OV series CMOS chips. The realization of SCCB on ARM9 embedded platform is emphasized. S3C2440 provides a camera interface, and the camera driver is designed based on it. Then the process of the application is introduced. The experiences show that the driver works well for video previewing and coding. The video quality can satisfy the requirements. The driver can be easily applicant on other embedded equipment.

Present an anti-geometric attacks video watermarking scheme based on normalization of pseudo-Zernike moment. Employing the RST invariant property of improved Pseudo-Zernike moment magnitudes, the watermark image is embedded into the centre circle area of I frame in the spatial domain after reconstruction. Because of much high similarity between video frames, an optimal moment set is selected on the basis of I frame and the adjacent P frame. In detection, the improved Pseudo-Zernike moments of the possibly attacked I frame are computed first and the ones of I frame are substituted by that of the adjacent P frame selected in the optimal moment set as the key to extract watermark. Experimental results show this scheme can get good quality images, improve the algorithm’s dependence on original data, and makes the watermark more robustness to geometric attacks, especially for scale attacks.

In view of the current status quo, which the feature- based digital watermarking is lack of robustness to non uniform scaling attacks, this paper puts forward a watermarking algorithm based template and local feature. Firstly, the template information is embedded in spatial domain. Then the improved Harris-Laplacian detector is utilized to extract steady feature points from the host image ,the local feature regions areas are constructed adaptively according to the feature scale theory; finally, the digital watermark are embedded into in the local feature regions using cirque quantization .Experimental results show that the proposed scheme is not only invisible and robust against common signals processing., but also robust against the geometric attacks In particular, non uniform scaling, etc. Keywords-geometric attacks; Harris-Laplacian; local feature regions