Signature Detection and Matching for Document Image Retrieval

As one of the most pervasive methods of individual identification and document authentication, signatures present convincing evidence and provide an important form of indexing for effective document image processing and retrieval in a broad range of applications. However, detection and segmentation of free-form objects such as signatures from clustered background is currently an open document analysis problem. In this paper, we focus on two fundamental problems in signature-based document image retrieval. First, we propose a novel multiscale approach to jointly detecting and segmenting signatures from document images. Rather than focusing on local features that typically have large variations, our approach captures the structural saliency using a signature production model and computes the dynamic curvature of 2D contour fragments over multiple scales. This detection framework is general and computationally tractable. Second, we treat the problem of signature retrieval in the unconstrained setting of translation, scale, and rotation invariant nonrigid shape matching. We propose two novel measures of shape dissimilarity based on anisotropic scaling and registration residual error and present a supervised learning framework for combining complementary shape information from different dissimilarity metrics using LDA. We quantitatively study state-of-the-art shape representations, shape matching algorithms, measures of dissimilarity, and the use of multiple instances as query in document image retrieval. We further demonstrate our matching techniques in offline signature verification. Extensive experiments using large real-world collections of English and Arabic machine-printed and handwritten documents demonstrate the excellent performance of our approaches.     

The choice of vantage objects for image retrieval

Suppose that we have a matrix of dissimilarities between n images of a database. For a new image, we would like to select the most similar image of our database. Because it may be too expensive to compute the dissimilarities for the new object to all images of our database, we want to find p?n “vantage objects” (Pattern Recognition 35 (2002) 69) from our database in order to select a matching image according to the least Euclidean distance between the vector of dissimilarities between the new image and the vantage objects and the corresponding vector for the images of the database. In this paper, we treat the choice of suitable vantage objects. We suggest a loss measure to assess the quality of a set of vantage objects: For every image, we select a matching image from the remaining images of the database by use of the vantage set, and we average the resulting dissimilarities. We compare two classes of choice strategies: The first one is based on a stepwise forward selection of vantage objects to optimize the loss measure. The second is to choose objects as representative as possible for the whole range of the database.     

Face recognition based on 3D ridge images obtained from range data

In this paper, we present an approach for 3D face recognition from frontal range data based on the ridge lines on the surface of the face. We use the principal curvature, k_max, to represent the face image as a 3D binary image called ridge image. The ridge image shows the locations of the ridge points around the important facial regions on the face (i.e., the eyes, the nose, and the mouth). We utilized the robust Hausdorff distance and the iterative closest points (ICP) for matching the ridge image of a given probe image to the ridge images of the facial images in the gallery. To evaluate the performance of our approach for 3D face recognition, we performed experiments on GavabDB face database (a small size database) and Face Recognition Grand Challenge V2.0 (a large size database). The results of the experiments show that the ridge lines have great capability for 3D face recognition. In addition, we found that as long as the size of the database is small, the performance of the ICP-based matching and the robust Hausdorff matching are comparable. But, when the size of the database increases, ICP-based matching outperforms the robust Hausdorff matching technique.     

Object recognition by three-dimensional curve matching

Experimental results for the recognition of general curves in three-space using registered range and intensity images are presented. the matching algorithm uses fast Fourier transforms to determine the least-squares difference between sequences of points sampled at equal intervals along two piecewise linear approximations of curves in three-space and returns the rotation and translation required to bring one of the curves into closest juxtaposition with the other. Performance of the algorithm is demonstrated by matching curves in three-space which are the boundaries of regions of contrasting reflectivity on curved surfaces. the experiments use a recently developed range sensor which is able to generate a 512 × 460 × 12 bit range image (with registered intensity image) in 40 s.     

Optimizing image matches via a verification model

In the literature, we have seen a boom in wide‐baseline matching approaches proposed for locating correspondences between images. However, wrong correspondences or the so‐called outliers are still rather inevitable, especially in urban environments with the presence of repetitive structures, and/or a large dissimilarity in viewpoints. In this paper, we propose a verification model to optimize the image matching results by significantly reducing the number of outliers. Several geometric and appearance‐based measurements are exploited, and conditional probability is used to compute the probability of each true correspondence. The model is validated by extensive experiments on images from the ZuBud database, which are taken in different weather conditions, seasons, and with different cameras. It is also demonstrated on a real‐time application of an image‐based navigation system. © 2010 Wiley Periodicals, Inc.     

An effective approach for iris recognition using phase-based image matching

This paper presents an efficient algorithm for iris recognition using phase-based image matching --- an image matching technique using phase components in 2D Discrete Fourier Transforms (DFTs) of given images. Experimental evaluation using CASIA iris image databases (ver. 1.0 and ver. 2.0) and Iris Challenge Evaluation (ICE) 2005 database clearly demonstrates that the use of phase components of iris images makes possible to achieve highly accurate iris recognition with a simple matching algorithm. This paper also discusses major implementation issues of our algorithm. In order to reduce the size of iris data and to prevent the visibility of iris images, we introduce the idea of 2D Fourier Phase Code (FPC) for representing iris information. 2D FPC is particularly useful for implementing compact iris recognition devices using state-of-the-art DSP (Digital Signal Processing) technology.     

Neighborhood preserving regression for image retrieval

In content-based image retrieval (CBIR), relevance feedback has been proven to be a powerful tool for bridging the gap between low level visual features and high level semantic concepts. Traditionally, relevance feedback driven CBIR is often considered as a supervised learning problem where the user provided feedbacks are used to learn a distance metric or classification function. However, CBIR is intrinsically a semi-supervised learning problem in which the testing samples (images in the database) are present during the learning process. Moreover, when there are no sufficient feedbacks, these methods may suffer from the overfitting problem. In this paper, we propose a novel neighborhood preserving regression algorithm which makes efficient use of both labeled and unlabeled images. By using the unlabeled images, the geometrical structure of the image space can be incorporated into the learning system through a regularizer. Specifically, from all the functions which minimize the empirical loss on the labeled images, we select the one which best preserves the local neighborhood structure of the image space. In this way, our method can obtain a regression function which respects both semantic and geometrical structures of the image database. We present experimental evidence suggesting that our algorithm is able to use unlabeled data effectively for image retrieval.     

Efficient planar object tracking and parameter estimation usingcompactly represented cubic B-spline curves

In this paper, we consider the problem of matching 2D planar object curves from a database, and tracking moving object curves through an image sequence. The first part of the paper describes a curve data compression method using B-spline curve approximation. We present a new constrained active B-spline curve model based on the minimum mean square error (MMSE) criterion, and an iterative algorithm for selecting the “best” segment border points for each B-spline curve. The second part of the paper describes a method for simultaneous object tracking and affine parameter estimation using the approximate curves and profiles. We propose a novel B-spline point assignment algorithm which incorporates the significant corners for interpolating corresponding points on the two curves to be compared. A gradient-based algorithm is presented for simultaneously tracking object curves, and estimating the associated translation, rotation and scaling parameters. The performance of each proposed method is evaluated using still images and image sequences containing simple objects     

非相似度保持投影

由主成分分析(PCA)可知任何一幅人脸图像都可以通过一组特征脸的线性加权来重构,PCA是最小均方误差意义下图像的最优表示,但是传统的PCA最终只通过比较加权系数的欧氏距离来进行识别,没有考虑残差。因此,提出非相似尺度的概念,将两个样本同时投影到相同向量上,在确定它们关系时既考虑投影系数,也考虑重构所产生的残差。两者的投影系数和残差相差越大,说明这两个样本越不相似。和保局投影(LPP)有所不同,非相似度保持投影算法不必预先设定近邻个数,它是利用非相似度的概念,创建非相似度散布矩阵,最终通过最大化目标函数获取最优子空间。在AR库和Feret库上的实验结果证明了该方法的有效性。     

A two‐stage sampling for robust feature matching

For any visual feature‐based SLAM (simultaneous localization and mapping) solutions, to estimate the relative camera motion between two images, it is necessary to find “correct” correspondence between features extracted from those images. Given a set of feature correspondents, one can use a n‐point algorithm with robust estimation method, to produce the best estimate to the relative camera pose. The accuracy of a motion estimate is heavily dependent on the accuracy of the feature correspondence. Such a dependency is even more significant when features are extracted from the images of the scenes with drastic changes in viewpoints and illuminations and presence of occlusions. To make a feature matching robust to such challenging scenes, we propose a new feature matching method that incrementally chooses a five pairs of matched features for a full DoF (degree of freedom) camera motion estimation. In particular, at the first stage, we use our 2‐point algorithm to estimate a camera motion and, at the second stage, use this estimated motion to choose three more matched features. In addition, we use, instead of the epipolar constraint, a planar constraint for more accurate outlier rejection. With this set of five matching features, we estimate a full DoF camera motion with scale ambiguity. Through the experiments with three, real‐world data sets, our method demonstrates its effectiveness and robustness by successfully matching features (1) from the images of a night market where presence of frequent occlusions and varying illuminations, (2) from the images of a night market taken by a handheld camera and by the Google street view, and (3) from the images of a same location taken daytime and nighttime.     

Stereo reconstruction from multiperspective panoramas

A new approach to computing a panoramic (360 degrees) depth map is presented in this paper. Our approach uses a large collection of images taken by a camera whose motion has been constrained to planar concentric circles. We resample regular perspective images to produce a set of multiperspective panoramas and then compute depth maps directly from these resampled panoramas. Our panoramas sample uniformly in three dimensions: rotation angle, inverse radial distance, and vertical elevation. The use of multiperspective panoramas eliminates the limited overlap present in the original input images and, thus, problems as in conventional multibaseline stereo can be avoided. Our approach differs from stereo matching of single-perspective panoramic images taken from different locations, where the epipolar constraints are sine curves. For our multiperspective panoramas, the epipolar geometry, to the first order approximation, consists of horizontal lines. Therefore, any traditional stereo algorithm can be applied to multiperspective panoramas with little modification. In this paper, we describe two reconstruction algorithms. The first is a cylinder sweep algorithm that uses a small number of resampled multiperspective panoramas to obtain dense 3D reconstruction. The second algorithm, in contrast, uses a large number of multiperspective panoramas and takes advantage of the approximate horizontal epipolar geometry inherent in multiperspective panoramas. It comprises a novel and efficient 1D multibaseline matching technique, followed by tensor voting to extract the depth surface. Experiments show that our algorithms are capable of producing comparable high quality depth maps which can be used for applications such as view interpolation.     

Utilizing similarity relationships among existing data for high accuracy processing of content-based image retrieval

Retrieving similar images based on its visual content is an important yet difficult problem. We propose in this paper a new method to improve the accuracy of content-based image retrieval systems. Typically, given a query image, existing retrieval methods return a ranked list based on the similarity scores between the query and individual images in the database. Our method goes further by relying on an analysis of the underlying connections among individual images in the database to improve this list. Initially, we consider each image in the database as a query and use an existing baseline method to search for its likely similar images. Then, the database is modeled as a graph where images are nodes and connections among possibly similar images are edges. Next, we introduce an algorithm to split this graph into stronger subgraphs, based on our notion of graph’s strength, so that images in each subgraph are expected to be truly similar to each other. We create for each subgraph a structure called integrated image which contains the visual features of all images in the subgraph. At query time, we compute the similarity scores not only between the query and individual database images but also between the query and the integrated images. The final similarity score of a database image is computed based on both its individual score and the score of the integrated image that it belongs to. This leads effectively to a re-ranking of the retrieved images. We evaluate our method on a common image retrieval benchmark and demonstrate a significant improvement over the traditional bag-of-words retrieval model.     

一种快速的子序列匹配算法

提出一种在时间序列上快速匹配子序列的算法,该算法不同于FRM算法,而是采用VA-file这种索引结构,将数据点直接存储在索引上,并在该索引的基础上设计了一种进行范围查询的方法.实验采用了三种时间序列数据集,从不同的角度验证算法的有效性,结果表明该算法大大提高了查询性能.     

Robust memory-efficient data level information fusion of multi-modal biometric images

This paper presents a novel multi-level wavelet based fusion algorithm that combines information from fingerprint, face, iris, and signature images of an individual into a single composite image. The proposed approach reduces the memory size, increases the recognition accuracy using multi-modal biometric features, and withstands common attacks such as smoothing, cropping, JPEG 2000, and filtering due to tampering. The fusion algorithm is validated using the verification algorithms we developed, existing algorithms, and commercial algorithm. In addition to our multi-modal database, experiments are also performed on other well known databases such as FERET face database and CASIA iris database. The effectiveness of the fusion algorithm is experimentally validated by computing the matching scores and the equal error rates before fusion, after reconstruction of biometric images, and when the composite fused image is subjected to both frequency and geometric attacks. The results show that the fusion process reduced the memory required for storing the multi-modal images by 75%. The integrity of biometric features and the recognition performance of the resulting composite fused image is not affected significantly. The complexity of the fusion and the reconstruction algorithms is O(n log n) and is suitable for many real-time applications. We also propose a multi-modal biometric algorithm that further reduces the equal error rate compared to individual biometric images.     

A graph-based optimization algorithm for fragmented image reassembly

We propose a graph-based optimization framework for automatic 2D image fragment reassembly. First, we compute the potential matching between each pair of the image fragments based on their geometry and color. After that, a novel multi-piece matching algorithm is proposed to reassemble the overall image fragments. Finally, the reassembly result is refined by applying the graph optimization algorithm. We perform experiments to evaluate our algorithm on multiple torn real-world images, and demonstrate the robustness of this new assembly framework outperforms the existing algorithms in both reassembly accuracy (in handling accumulated pairwise matching error) and robustness (in handling small image fragments).     

基于中心点的指纹细节结构匹配算法

指纹细节匹配算法是自动指纹识别系统(AFIS)中一项关键的任务,目前存在大量的研究和算法.依据算法是否依赖中心点,指纹细节点匹配算法可以分为两类:基于中心点的匹配算法和非中心点匹配算法.大多数非中心点匹配算法都非常耗时,因此不适合在线应用.而基于中心点方法的效率相对较高,但是这类算法极度依赖于中心点的定位精度.在本文中,提出了一种全新的基于中心点的指纹细节结构匹配算法,该算法综合了基于中心点匹配算法和非中心点匹配算法的优点,同时又避免了二者的缺点.首先利用中心点检测算法获得中心点的位置,然后在中心区域定义了一些局部的结构,同时利用这些局部结构寻找指纹细节的对应点,并通过对应点和中心点的相对关系来确认这些对应细节点.其次利用这些细节对应点匹配全局的细节信息,最后,利用匹配细节的全局距离和距离方差来判决最终匹配结果.实验结果表明,算法的匹配效果非常好,同时匹配效率较高,非常适合在线指纹识别系统的应用.     

由整体到局部的平面曲线部分匹配算法

在基于曲线匹配的检索系统中,提高曲线的匹配速度和精度具有重要的意义.提出一种平面曲线的部分匹配算法,该算法分为整体搜索和局部匹配2个阶段.首先整体搜索确定候选的匹配区域,然后在局部进行精确匹配和验证.对于特征点较少的曲线,根据曲率极值点将曲线划分为多条曲线段,采用局部线性搜索法实现曲线的部分匹配.实验结果说明了算法的有效性.     

An algorithm for the visualization of stereo images simultaneously captured with different exposures

The visualization of stereo images obtained from two eyepiece cameras of a stereo microscope with different exposures is studied. This problem is solved to improve the quality of resulting images in the case when one image is not sufficient for capturing an object with the desired color reproduction accuracy and high level of detail both in dark and light regions. An approach to solving this problem in which differently exposed images are split between two views is considered. This approach allows us to significantly reduce the capturing time and to enhance the quality of capturing moving objects. The algorithm described in [1] is used as the basic algorithm; the main steps of this algorithm are the stereo matching of two input images and the construction of high dynamic range images. Modifications of the basic algorithm that use different stereo matching techniques are proposed. The application of the algorithm described in [2] for the visualization of stereo images without constructing high dynamic range images is discussed. A database of images captured with different exposures by a stereo microscope is created. The quality of algorithms applied to the images from this database is evaluated in the HDR-VDP-2.2 metric [3].