Catálogo de publicaciones - libros

Image steganography differs from cryptography that directly encrypts the data transmitted, embeds important messages into a digital image that appears usual before transmission to improve the concealment of communication. It proposes a novel and secure steganographic method for embedding secret data into a gray scale image. It first partitions the cover image into blocks of the same size and calculates the mean of all pixels in each blocks. And then embed different amount of data bits into different pixels in a block according to the range of the difference value between each pixels and the block mean. In this way, the proposed method different amount of data into blocks of different degree of smoothness. Experimental results show that the quality of both the extracted important image and the stego-image are acceptable. And have the successful way of embedding secret data into an image without noticeable distortion.

Virtual-environment-based instrument development system (VEBIDS) is a concept previously put forward by the author to develop instruments in a virtual environment (VE) generated by computer, with the support of appropriate computer hardware, software, virtual reality (VR) peripherals in accordance with specific applications. This paper first presents the integrated model of VEBIDS, then discusses the application of case-based reasoning (CBR) in it. To make the system more efficient and flexible, artificial neural network (ANN) is also adopted into the system to complement with CBR. Experimental results show that the combination of CBR and ANN in VEBIDS is satisfactory. With the appearance of more economical and user-affordable high-performance computers and VR peripherals, VEBIDS can become a very practical and useful tool in instrument development.

In this paper, we propose a real-time video watermarking chip and system which is hardware based watermark embedding system of SD/HD (standard definition/high definition) video with STRATIX FPGA device from ALTERA as well as we implement the watermark detector which is hardware and software based detector system with STRATIX FPGA, PCI controller and VC++. There was little visual artifact due to watermarking in subjective quality evaluation between the original video and the watermarked one. Embedded watermark was all extracted after a robustness test called natural video attacks such as A/D (analog/digital) conversion and MPEG compression. Our implemented watermarking hardware system can be useful in movie production and broadcasting companies that requires real-time based copyright protection system.

The most widely used methods for time- frequency analysis belong to Cohen’s class of spectrogram estimators. Wigner–Ville distribution (WVD) and multiwindow time -frequency analysis are the most popular techniques. One serious limitation of Wigner distribution is that it produces cross-terms of large magnitudes. There are some limitations to multiwindow time -frequency analysis also. In this paper, a spectrogram using Gaussian–Hermite window is presented. Results indicate that this spectrogram produces cross-terms of very small magnitudes which can be eliminated by thresholding. This spectrogram also performs well where multiwindow analysis fails. The technique has lower computational complexity as compared to multiwindow spectral analysis and other methods for joint time -frequency distributions.

Ordinary high-speed block matching algorithms have a disadvantage that they need to get MAD (Mean Absolute Distance) as many as the number of search points due to comparing the MAD between the current frame’s search block and the reference frame’s search block. To solve such disadvantage of high-speed block matching algorithm, the proposed high-speed block matching algorithm employs a MMAD calculation method using a specific characteristic that neighboring pixels have almost same values. In this paper, we can get rid of unnecessary MAD calculation between the search point blocks by the new calculation method which uses the previously calculated MAD as the current search point and by breaking from the established MAD calculation method which calculates the MAD of a new search point by each search stage. Comparing with the established high-speed block matching algorithm, this new calculation’s estimated motion error was shown as similar, and the total calculation amount decreased by 2FN2Ep.

Multiwavelet analysis has been a powerful tool in image processing. But the theory is seldom applied to denoise cellular images of common phytoplankton. This paper proposes a new method which uses multiwavelets combining soft thresholding to meet this kind of applications. Three types of mutiwavelets thresholding are considered: scalar, decor and vector. Results of numerical experiments show that this method particularly using Chui-Lian orthonormal multiwavelet has better effectiveness than other methods referred in this paper for removing Gaussian noise.

This paper proposes one aerial target recognition algorithm based on moment characteristics and immunity computation. The algorithm first uses the thresholding segmentation technology based on immunity computation to make target extraction, and calculates its moment characteristics. Then it uses a neural network identifier of evolution weight to carry on the target recognition and determine final target category. A simulation experiment indicates that the method based on moment characteristics and immunity computation can enhance the recognition efficiency and has achieved good results.

The mean shift algorithm has achieved considerable success in object tracking due to its simplicity and robustness. It finds local minima of a similarity measure between the color histograms or kernel density estimates of the target and candidate image. The most typically used similarity measures are the Bhattacharyya coefficient. There are assumptions which limited lighting condition due to color is very sensitive about illuminations. And the algorithm has weakness about inference of another object. In this paper we propose method that combined advantage of color distribution and depth. As apply robust error norm, problems are conquer. The method is useful for face tracking under the dynamic illumination. Also it voids an interference of another object.

Connecting computational results with biological knowledge is one of the most urgent issues in the post genome research. We present a simple but powerful procedure to identify co-expressed gene groups associated with significant functional categories. An R language implementation, named SigClust, was built and tested on a set of gene expression dataset. Further, comparison of linear and nonlinear similarity measurements of gene expression profiles was tested on SigClust. Finally, a consistent conclusion with previous researches was acquired.

An efficient approach based on particle swarm clustering is proposed for fish bone detection. With respect to the gray value distribution of the radiographic image, Gaussian distribution properties are used combined with morphological methods for image preprocessing, by which one region of interest (ROI) is obtained. Then, particle swarm clustering is used to segment the ROI into different clusters. Finally, experiments are done and comparisons with traditional image segmentation techniques show that our approach is effective.