This work proposes an improved reversible data hiding scheme in encrypted images using parametric binary tree labeling(IPBTL-RDHEI), which takes advantage of the spatial correlation entire original image but not small blocks to reserve room for data. Then is with encryption key and used label pixels into two different categories. Finally, one categories can embed secret information by bit replacement. According experimental results, compared several state-of-the-art methods, proposed...

Spatial re-ranking has proved to be successful in image retrieval. Yet no work on spatial been systematically reported for instance search from videos so far: As are composed of multiple frames, and frame-by-frame verification is too prohibitive, till now we lack an efficient method designed the purpose video search. The effectiveness unclear as well. This paper proposes a practical We make two contributions speed up algorithm: One select most representative image/frame verification, based...

It is widely recognized that clustering ensemble fit for any shape and distribution dataset the boosting method provides superior results classification problems. In paper, a dual proposed fuzzy . At each iteration, new training set created based on original datasets' probability which associated with previous clustering. According to method, subset contains not only instances hard cluster in stages , but also easy cluster. The final solution produced by using co-association matrix....

A data hiding method based on a reference matrix generated base-5 numeral system was presented in this article. The fully exploits the embedding space to achieve high capability and acceptable quality. results of experiment indicate that average value PSNR (Peak Signal Noise Ratio) stego images with full payload by is about 43.7dB, which demonstrates good visual Furthermore, an capacity 2.3219 bpp (bit per pixel) achieved proposed method, means for image sized 512×512, 608679 bits can be...

A robust and fast image segmentation method is provided based on color distribution. In this paper, unlike the widely used training optimizing process, an recursively divided only according to differences, a whole region into two regions that have most difference Gaussian Mixture Model distribution in space, segmented separate continuous after few iterations. Morphological post process introduced get correct segment results with closed semantical boundary. At last of experimental comparison...

In order to improve the convergence rate of genetic algorithm based on edge detection, a novel detection method good point set algorithm(GGA) was proposed. The proposed first redesigns crossover operation by using theory in which progeny inherits common genes parents represent its family so as algorithm. Furthermore, offers another better way rate, that is, reduce solution domain pre-processing image filtering non pixel before executing. Experimental results show performs very well terms...

Clustering ensemble is fit for any shape and distribution datset . Boosting methodogy provides superior results classification problems. In the paper, A dual boosting proposed of fuzzy clustering At iteration , a new training set created based on original datasets' weights which associated with previous According method, not only include datas hard to ,but also includes dta easy cluster The final solution propuced by re-clustering co-association matrix. Experiments both artifical real word...

In this paper, a novel watermarking scheme based on relationship of Tchebichef Moments is proposed. Firstly, the image divided into blocks; then moment's each block computed, which computed rapidly with help look-up-table; finally, moments are modified to embed watermark. The watermark extraction reverse procedure embedding and original not necessary in detection. Experimental results demonstrate that proposed technique robustness attacks such as JPEG compress, noises, luminance contrast...

In this paper, An improved resources search algorithm in Gnutella network is proposed by analyzing and summarizing current related algorithm. Experiments showed that can more effectively reduce traffic than flooded policy system. improve P2P network's integrity, availability scalability.

In this paper, we summarize the characteristics of message search mechanism Gnutella P2P network model. By analyzing its disadvantages, an improvement resource discovery algorithm is proposed, it can more effectively reduce traffic than existing methods in system, and also improve network's integrity, availability scalability.

The local feature (e.g. SIFT) and Bag of Words (BOW) model play key roles in achieving a state-of-the-art performance for image classification. Although we realize that utilizing extra color information will undoubtedly boost the feature, there still have not been any research carefully focused on how to efficiently transfer this boosted into BOW. In paper, channel-wise separate quantization is newly proposed replace traditional multi-channel combined SIFT. Dozens experiments conducted...

Image semantic has the characters of vague and abstractive, therefore, only low-level feature is not enough to describe image semantics. It requires combine with related content in order improve accuracy annotation In this paper, An algorithm based on mutual K adjacency graph proposed which describes relationship between features, annotated words image. Semantic information extracted from paired-nodes graph, improves effectively performance. Based analysis structure Combined Random Walk...