In the computation process of many kernel methods, one of the important step is the formation of the kernel matrix. But the size of kernel matrix scales with the number of data set, it is infeasible to store and compute the kernel matrix when faced with the large-scale data set. To overcome computational and storage problem for large-scale data set, a new framework, matrix-based kernel method, is proposed. By initially dividing the large scale data set into small subsets, we could treat the autocorrelation matrix of each subset as the special computational unit. A novel polynomial-matrix kernel function is then adopted to compute the similarity between the data matrices in place of vectors. The proposed method can greatly reduce the size of kernel matrix, which makes its computation possible. The effectiveness is demonstrated by the experimental results on the artificial and real data set.

This paper proposes a new theory of adder and its basic structure. The new adder of asynchronous structure constructed by half adders, called Parallel Feedback Carry Adder (PFCA) as its carry mode is parallel feedback. In theory, the area consumption of n-bit PFCA is close to O(n) and the average length of carry chain is O(log n). A CMOS gate implementation scheme is implemented. HSPICE simulation results show that PFCA has obvious advantages over RCA, CLA, CSeA in speed and area, especially when n is bigger.

Original models for direction relations ignored the restriction of topology and distance relations to direction representation. To improve the representation of direction relations model by pondering about the influence of topology and distance relations on direction relations, we categorize direction reference frame into topological reference and coarse directions reference and present a new direction relations quantitative and statistics models based on the new direction reference frames. Instead of degree, this new model uses a coordinate-based quantitative method to describe direction relations for the distance restrain, while it reflects the constraints of topology by the direction reference frame and by the coordinate representation. It covers all intricacies imposed by different types of objects and has more sensitivity to the configuration of objects. Experiments have been carried out and the results indicate the excellent efficiency in view of directional description.

Leaf features play an important role in plant species identification and plant taxonomy. The type of the leaf vein is an important morphological feature of the leaf in botany. Leaf vein should be extracted from the leaf in the image before discriminating its type. In this paper a new method of leaf vein extraction has been proposed based on gray-scale morphology. Firstly, the color image of the plant leaf is transformed to the gray image according to the hue and intensity information. Secondly, the gray-scale morphology processing is applied to the image to eliminate the color overlap in the whole leaf vein and the whole background. Thirdly, the linear intensity adjustment is adopted to enlarge the gray value difference between the leaf vein and its background. Fourthly, calculate a threshold with OSTU method to segment the leaf vein from its background. Finally, the leaf vein can be got after some processing on details. Experiments have been conducted with several images. The results show the effectiveness of the method. The idea of the method is also applicable to other linear objects extraction.

An improved local equilibrium contrast enhancement algorithm based self-adaptive contrast enhancement algorithm is proposed for infrared laser images, in which the image pixel value histogram is divided into three parts: background and noise area, targets area, and uninterested area. The targets parts are highlighted, while the background and noise parts and the uninterested parts are restrained. A comprehensive qualitative and quantitative image enhancement performance evaluation is presented to verify the proposed theory and algorithm validity, efficiency and reasonability. The experimental results indicate that the proposed algorithm can greatly improve the global and local contrast for both near infrared images and far infrared laser images while efficiently reducing noise in the infrared laser images，and the visual quality of enhanced image is obviously better than the enhancement of the traditional histogram equalization, double plateaus histogram equalization algorithm, etc.

Recently, chaos has attracted much attention in the field of cryptography. To study the security with a known image of a symmetric image encryption scheme, the attack algorithm of equivalent key is given. We give the known image attacks under different other conditions to obtain the equivalent key. The concrete step and complexity of the attack algorithm is given. So the symmetric image encryption scheme based on 3D chaotic cat maps is not secure.

Parameter quantization is very important for the synthetic speech quality of the vocoder. A new distortion measure for pitch as well as lsf quantization in ultra low bit rate Vocoder, whose parameters for several consecutive frames are grouped into a vector and jointly quantized to obtain high coding efficiency, is proposed based on mixed excitation linear prediction(MELP) vocoder. The product of sum of band pass voicing coefficients and gain parameter is used to denote the weighting factor of pitch as well as lsf parameters of current speech frame in the consecutive frames using weighted squared Euclidean distance measure to search the vector codebook. Comparing with the traditional method for a constant weighting factor by distinguishing Voiced/Unvoiced(UV) pattern of each speech frame, objective test results show that the quantization distortion of pitch is reduced by 3.3% and the mean opinion score (MOS) is increased by almost 0.1(3.5%).

Currently, edge detection is an effective means of collecting and analyzing various diseases information from mural collections by using this and data mining based on digital orthophoto map (DOM). But it is hard to extract better edges of mural diseases with traditional edge detection algorithms. Therefore, a new K-means Sobel algorithm is proposed and two evaluation factors are given to judge the extracting effect. Experiment results demonstrate that we can get a better effect by using new method than traditional algorithms. At last, vectorizing detected results, we can gain diseases areas. On that basis, a decision tree about mural diseases severities is established to provide useful information for mural diseases investigation and repair.