Longzhuang Li

Visual Segmentation-Based Data Record Extraction from Web Documents

Semi-structured data records contained in the Web pages provide useful information for shopping agents and metasearch engines. In this paper, we present a visual segmentation-based data record extraction (VSDR) method to extract data records from those Web pages. VSDR method first segments a Web page into semantic blocks using the spatial closeness and visual resemblance of data records, then neighboring and non-neighboring data records are extracted based on a compress and collapse technique. Experimental results slum that unlike the existing methods which only generate good results on their test domains, VSDR is a general data record extraction method that is able to produce quite stable and good results on a wide range of Web pages.

3D shape matching using collinearity constraint

In this paper, a novel algorithm is proposed to carry out automatic 3D shape matching with 3D shapes represented as sets of points. After the possible matches between the 3D shapes have been determined by the tradition ICP criterion, the novel approach employs the collinearity constraint to eliminate false matches based on a statistical model. A comparative study based on real range images has shown that the proposed algorithm is accurate, robust and efficient for the automatic matching of overlapping 3D shapes.

Incremental mesh-based integration of registered range images: robust to registration error and scanning noise

Existing integration algorithms often assume that the registration error of neighbouring views is an order of magnitude less than the measurement error [3]. This assumption is very restrictive that automatic registration algorithms can hardly meet. In this paper, we develop a novel integration algorithm, robust to both large registration errors and heavy scanning noise. Firstly, a pre-processing procedure is developed to automatically triangulate a single range image and remove noisy triangles. Secondly, we shift points along their orientations by the projection of their resulting correspondence vectors so that new correspondences can approach together, leading large registration errors to be compensated. Thirdly, overlapping areas between neighbouring views are detected and integrated, considering the confidence of triangles, which is a function of the including angle between the centroid point vector of a triangle and its normal vector. The outcome of integration is a set of disconnected triangles where gaps are caused by the removal of overlapping triangles with low confidence. Fourthly, the disconnected triangles are connected based on the principle of maximizing interior angles. Since the created triangular mesh is not necessarily smooth, finally, we minimize the weighted orientation variation. The experimental results based on real images show that the proposed algorithm significantly outperforms an existing algorithm and is robust to both registration error and scanning noise.

Projecting registration error for accurate registration of overlapping range images

Abstract In this paper, we propose a novel algorithm for the automatic registration of two overlapping range images. Since it is relatively difficult to compare the registration errors of different point matches, we project them onto a virtual image plane for more accurate comparison using the classical pin-hole perspective projection camera model. While the traditional ICP algorithm is more interested in the points in the second image close to the sphere centred at the transformed point, the novel algorithm is more interested in the points in the second image as collinear as possible to the transformed point. The novel algorithm then extracts useful information from both the registration error and projected error histograms for the elimination of false matches without any feature extraction, image segmentation or the requirement of motion estimation from outliers corrupted data and, thus, has an advantage of easy implementation. A comparative study based on real images captured under typical imaging conditions has shown that the novel algorithm produces good registration results.

Free form shape matching using deterministic annealing and softassign

Free form shape matching is a fundamental problem in both the machine vision and pattern recognition literatures. However, the automatic approach to free form shape matching still remains open. We improve an existing deterministic annealing and softassign method for automatic free form shape matching through incorporating rigid motion constraints. The experiments based on both synthetic data and real images show the effectiveness of the improved algorithm.

Qualitative Spatial Representation and Reasoning for Data Integration of Ocean Observational Systems

Spatial features are important properties with respect to data integration in many areas such as ocean observational information and environmental decision making. In order to address the needs of these applications, we have to represent and reason about the spatial relevance of various data sources. In the paper, we develop a qualitative spatial representation and reasoning framework to facilitate data retrieval and integration of spatial-related data from ocean observational systems, such as in situ observational stations in the Gulf of Mexico. In addition to adopt the state-of-the-art techniques to represent partonomic, distance, and topological relations, we develop a probability-based heuristic method to uniquely infer directional relations between indirectly connected points. The experimental results show that the proposed method can achieve the overall adjusted correct ratio of 87.7% by combining qualitative distance and directional relations.

Evaluating collinearity constraint for automatic range image registration

While most of the existing range image registration algorithms either have to extract and match structural (geometric or optical) features or have to estimate the motion parameters of interest from outliers corrupted point correspondence data for the elimination of false matches in the process of image registration, the registration error and the collinearity error derived directly from the traditional closest point criterion are also capable of doing the same job. However, the latter has an advantage of easy implementation. The purpose of this paper is to investigate which definition of collinearity is more accurate and stable in eliminating false matches inevitably introduced by the closest point criterion. The experiments based on real images show the advantages and disadvantages of different definitions of collinearity.

Qualitative spatial representation and reasoning for data integration of ocean observing systems

Abstract Spatial features are important properties with respect to data integration in many areas such as ocean observational information and environmental decision making. In order to address the needs of these applications, we have to represent and reason about the spatial relevance of various data sources to facilitate retrieval and integration of data. In this paper, using the in situ ocean observing stations in the Gulf of Mexico as an example we develop a statistical method, the semi-circular method, based on the semi-circular normal distribution to uniquely reason directional relations between indirectly connected points in addition to adopt the state-of-the-art qualitative spatial representation and reasoning techniques to represent partonomic, distance, and topological relations. In the experiment, the performance of the semi-circular method is compared with three existing methods, and the experimental results show that the statistic-based semi-circular method obtains the overall adjusted correct ratio of 88.1% by combining qualitative distance and directional relations, which achieves the comparable accuracy as and is slightly better than the probabilistic-based heuristic method.

Probabilistic discovery of motifs in water level

The discovery of water level time series motifs is of much importance to improve the water level predictions. These predictions thereby are useful to the shipping industry, people living in the coastal areas, and even for emergency evacuation in case of a hurricane. In this paper, symbolic aggregate approximation (SAX) is employed to index and reduce the dimension of the time series, and the random projection algorithm is used to discover the unknown time series motifs, which are tested for accuracy by comparing them with the brute force algorithm.

3D free form surface matching based on orientation difference length distribution

In this paper, we propose using orientation difference length distribution (ODLD) to represent a 3D view of each object with a free form surface. ODLD has a number of advantages over the shape distribution. A comparative study of ODLD and the shape distribution has shown that the former is significantly more accurate for 3D free form surface matching than the latter.