Locality-preserved maximum information projection.

Dimensionality reduction is usually involved in the domains of artificial intelligence and machine learning. Linear projection of features is of particular interest for dimensionality reduction since it is simple to calculate and analytically analyze. In this paper, we propose an essentially linear projection technique, called locality-preserved maximum information projection (LPMIP), to identify the underlying manifold structure of a data set. LPMIP considers both the within-locality and the between-locality in the processing of manifold learning. Equivalently, the goal of LPMIP is to preserve the local structure while maximize the out-of-locality (global) information of the samples simultaneously. Different from principal component analysis (PCA) that aims to preserve the global information and locality-preserving projections (LPPs) that is in favor of preserving the local structure of the data set, LPMIP seeks a tradeoff between the global and local structures, which is adjusted by a parameter alpha, so as to find a subspace that detects the intrinsic manifold structure for classification tasks. Computationally, by constructing the adjacency matrix, LPMIP is formulated as an eigenvalue problem. LPMIP yields orthogonal basis functions, and completely avoids the singularity problem as it exists in LPP. Further, we develop an efficient and stable LPMIP/QR algorithm for implementing LPMIP, especially, on high-dimensional data set. Theoretical analysis shows that conventional linear projection methods such as (weighted) PCA, maximum margin criterion (MMC), linear discriminant analysis (LDA), and LPP could be derived from the LPMIP framework by setting different graph models and constraints. Extensive experiments on face, digit, and facial expression recognition show the effectiveness of the proposed LPMIP method.     

Deriving evidence theoretical functions in multivariate data spaces: a systematic approach.

The mathematical theory of evidence is a generalization of the Bayesian theory of probability. It is one of the primary tools for knowledge representation and uncertainty and probabilistic reasoning and has found many applications. Using this theory to solve a specific problem is critically dependent on the availability of a mass function (or basic belief assignment). In this paper, we consider the important problem of how to systematically derive mass functions from the common multivariate data spaces and also the ensuing problem of how to compute the various forms of belief function efficiently. We also consider how such a systematic approach can be used in practical pattern recognition problems. More specifically, we propose a novel method in which a mass function can be systematically derived from multivariate data and present new methods that exploit the algebraic structure of a multivariate data space to compute various belief functions including the belief, plausibility, and commonality functions in polynomial-time. We further consider the use of commonality as an equality check. We also develop a plausibility-based classifier. Experiments show that the equality checker and the classifier are comparable to state-of-the-art algorithms.     

Structural-symmetry-related sequence patterns of the proteins of beta-propeller family

It is still not very clear to what extent and how does the amino acid sequences of proteins determine their tertiary structures. In this paper, we report our investigations of the sequence-structure relations of the proteins in the beta-propeller fold family, which adopt highly symmetrical tertiary structures while their sequences appear "random". We analyzed the amino acid sequences by using a similarity matrix plus Pearson correlation method and found that the sequences can show the same symmetries as their tertiary structures only if we deduce the conditions of sequence similarity. This suggests that some key residues may play an important role in the formation of the tertiary structures of these proteins.     

Batch Nearest Neighbor Search for Video Retrieval

To retrieve similar videos to a query clip from a large database, each video is often represented by a sequence of high- dimensional feature vectors. Typically, given a query video containing m feature vectors, an independent nearest neighbor (NN) search for each feature vector is often first performed. After completing all the NN searches, an overall similarity is then computed, i.e., a single content-based video retrieval usually involves m individual NN searches. Since normally nearby feature vectors in a video are similar, a large number of expensive random disk accesses are expected to repeatedly occur, which crucially affects the overall query performance. Batch nearest neighbor (BNN) search is stated as a batch operation that performs a number of individual NN searches. This paper presents a novel approach towards efficient high-dimensional BNN search called dynamic query ordering (DQO) for advanced optimizations of both I/O and CPU costs. Observing the overlapped candidates (or search space) of a pervious query may help to further reduce the candidate sets of subsequent queries, DQO aims at progressively finding a query order such that the common candidates among queries are fully utilized to maximally reduce the total number of candidates. Modelling the candidate set relationship of queries by a candidate overlapping graph (COG), DQO iteratively selects the next query to be executed based on its estimated pruning power to the rest of queries with the dynamically updated COG. Extensive experiments are conducted on real video datasets and show the significance of our BNN query processing strategy.     

Three-dimensional reconstruction on PC-Windows platform for evaluation of living donor nephrectomy

Evaluation of renal vasculature was necessary for preoperative donor assessment in living donor kidney transplantation, and the ability to view the vascular imaging in three-dimensional (3D) space should be helpful undoubtedly. Considering the widespread use of personal computer (PC) systems, we aimed to find a handy way to display the anatomy of the renal arterial and venous systems of potential donors on PC-Windows platform. Ten living-related donors were included in this study. Serial computed tomography (CT) images were loaded into Amira 3.1 running on a PC with Windows XP. Following image crop, segmentation and 3D reconstruction, we got the 3D images and the video clips. According to the displays which were confirmed by latter intraoperative findings, eight donors had single renal vessels, one had a left accessory renal artery and one had a right accessory renal artery. This project offered a new approach to evaluate the renal vessel anatomy in living donor kidney transplantation, and it was favorable for accuracy and popularizing.     

LSS: an S-Plus/R program for the accelerated failure time model to right censored data based on least-squares principle

Due to lack of proper inference procedure and software, the ordinary linear regression model is seldom used in practice for the analysis of right censored data. This paper presents an S-Plus/R program that implements a recently developed inference procedure [Z. Jin, D.Y. Lin, Z. Ying, On least-squares regression with censored data, Biometrika 93 (2006) 147-161] for the accelerated failure time model based on the least-squares principle. The program is user-friendly and yields outputs similar to S-Plus/R function 1m.     

Evaluation of amplitude-based sorting algorithm to reduce lung tumor blurring in PET images using 4D NCAT phantom

PURPOSE: develop and validate a PET sorting algorithm based on the respiratory amplitude to correct for abnormal respiratory cycles. METHOD AND MATERIALS: using the 4D NCAT phantom model, 3D PET images were simulated in lung and other structures at different times within a respiratory cycle and noise was added. To validate the amplitude binning algorithm, NCAT phantom was used to simulate one case of five different respiratory periods and another case of five respiratory periods alone with five respiratory amplitudes. Comparison was performed for gated and un-gated images and for the new amplitude binning algorithm with the time binning algorithm by calculating the mean number of counts in the ROI (region of interest). RESULTS: an average of 8.87+/-5.10% improvement was reported for total 16 tumors with different tumor sizes and different T/B (tumor to background) ratios using the new sorting algorithm. As both the T/B ratio and tumor size decreases, image degradation due to respiration increases. The greater benefit for smaller diameter tumor and lower T/B ratio indicates a potential improvement in detecting more problematic tumors.     

一种通用的数据类型——十进制浮点数

为智能仪表提供一种“十进制浮点数”,其特点是值域宽、值准确,它能代替整、长整型、三字节浮点数、四字节浮点数、BCD码数。用它通信可简化和统一仪表的通信协议,为制定仪表通用通信协议标准创造条件。