A hotel recommendation system based on collaborative filtering method of clustering and Rankboost algorithm proposed in this paper, which can avoid the cold-start scalability problems existing traditional filtering. One find a quickly efficiently when he uses system.

Static analysis is a crucial protection layer that enables modern antivirus systems to address the rampant proliferation of malware. These are increasingly relying on deep neural networks (DNNs) automatically extract reliable features and achieve outstanding detection accuracy. Since DNNs known be vulnerable adversarial examples, several studies have proposed practical evasion attacks generate perturbations can evade malware detectors. attacks, however, require specific designs for given...

We propose a background subtraction method for moving cameras based on trajectory classification, image segmentation and label inference.In the classification process, PCA-based outlier detection strategy is used to remove outliers in foreground trajectories.Combining optical flow with watershed algorithm, we trajectory-controlled algorithm which effectively improves edge-preserving performance prevents over-smooth problem.Finally, inference Markov Random field conducted labeling unlabeled...

Modern camera lenses become increasingly more complex to optimize the light efficiency of optical system. Recent research has combined single lens optics with post-capture correction methods based on computational photography. This study further improves design by correcting chromatic aberrations, after which a simple image deconvolution method is sufficient produce high-quality image. We initially estimate point spread function blind method. add Gaussian regularization as kernel prior...

In recent years, a lot of research has been conducted in the field object detection for aerial image, and many effective methods have emerged. This paper enhances existing Faster RCNN model to achieve better accuracy. We improve feature extraction using multi-scale fusion. Compared with traditional model, accuracy on mAP is improved by 1.06% applying proposed method. The visual effects numerical results verify improvement

Accurate workload prediction plays a key role in intelligent scheduling decisions on cloud platforms. There are massive amounts of short-workload sequences the platform, and small amount data presence outliers make accurate sequence challenge. For above issues, this paper proposes an ensemble learning method based sample weight transfer long short-term memory (LSTM), termed as Tr-Predictor. Specifically, selection similar combining time warp edit distance (TWED) entropy (TE) is proposed to...

With the development of computational photography, single-lens camera combined with corresponding image deblurring algorithm is gradually becoming a new research direction, replacing complex modern optical imaging system such as single lens reflex (SLR) camera. For camera, Point Spread Function (PSF) estimation accuracy will directly affect restoration effect. In this paper, we designed simple-lens cameras one, two and three lenses, respectively, propose robust accurate PSF method The key...

A novel object detection approach based on variance analysis is presented in this paper. Firstly, the of frame difference information analyzed and background model built. Then, used as threshold to separate image into two parts, moving region region. Thirdly, further by comparing it with background; uncovering are detected. Lastly, updated area for detecting next video frame. The experiment results show that proposed method practical fast speed, could detect clear motion objects meet need...

Batch mode active learning, where a batch of samples is simultaneously selected and labeled, challenging task. The challenge lies in how to maintain the informativeness keep diversity concurrently. We propose novel learning that balances representativeness using multi-set clustering. Our method utilizes sequential learner retain by providing ranking unlabeled constructing multiple informative sets for subsequent clusterings. K-means clustering used minimize redundancy among these improve...

This paper proposes a novel video stitching method that improves real-time performance and visual quality of multi-camera surveillance system.A two-stage seam searching algorithm based on enhanced dynamic programming is proposed.It can obtain satisfactory result achieve better than traditional seam-searching methods.The experiments show the computing time reduced by 66.4% using proposed compared with programming, while accuracy maintained.A local update scheme reduces deformation effect...

Fisheye imaging technique plays an important role in various areas. However, fisheye lenses can cause severe image distortion, which requires distortion rectification for these images. after rectification, the images will suffer from problem of resolution reduction, and degree reduction varies different regions. This directly affect subsequent process content analysis. Therefore, it is needed to adopt a super-resolution reconstruction algorithm enhance rectified image. paper designs...

This study proposes a data separation algorithm with computationally efficient strategies for non-blind deconvolution in the frequency domain. First, blurred image is separated into couple of basis and corresponding coefficients. Then traditional method domain employed pre-processing step deblurred bases are saved. For new same blur kernel, iterative optimization converted linear addition multiplication operations. Results qualitative quantitative evaluations demonstrate efficiency...

Region of interest extraction is very important in video analysis and understanding. A simple fast region method for sport scene images proposed this paper. Firstly, a applied to detect the pixels by defined function strategy improve computation speed mean each pixel; Secondly, all detected neighbored other are grouped object regions; after that, non-interest enclosed set as inner filling method; Thirdly, binarized composed can be easily segmented applying edge detection algorithms or...

In ALV system based on stereo vision, the obstacle detection is one of most important problems. this paper, a matching algorithm proposed region invariant moment to detect in and several novel strategies are introduced during object extraction stage. Firstly, simple pre-processing method binarize images by defined binarization function; Secondly, inner filled segmented using classic edge or outer contour tracing regions left right detected respectively; Lastly, matched between image moment,...

Abstract Infrastructure networks are critical components of contemporary society, and numerous approaches have been suggested for the selection strategies to protect these networks. However, uncertain environments, research on attack defense game models infrastructure is limited. Therefore, after reviewing existing approaches, a method based interval-valued intuitionistic fuzzy set (IVIFS) theory proposed games in First, we present process constructing model this paper, which mainly includes...

Recently, research interest in the field of infrastructure attack and defense scenarios has increased. Numerous methods have been proposed for studying strategy interactions that combine complex network theory game theory. However, unavoidable effect constrained strategies situations not considered previous studies. This study introduces a novel approach to analyzing these by including effects strategies, factor often neglected traditional analyses. First, we introduce rule constraints on...

The field of infrastructure security has garnered significant research attention. By integrating complex network theory with game theory, researchers have proposed many methods for studying the interactions between attacker and defender from a macroscopic viewpoint. We constructed model networks to analyze attacker-defender confrontations. To address challenge finding Nash equilibrium, we developed novel algorithm—node-incremental greedy algorithm (NIGA)—which uses less strategy space solve...