Similarity Measure For Matching Research Articles

Research on Rejoining Bone Stick Fragment Images: A Method Based on Multi-Scale Feature Fusion Siamese Network Guided by Edge Contour

The rejoining of bone sticks holds significant importance in studying the historical and cultural aspects of the Han Dynasty. Currently, the rejoining work of bone inscriptions heavily relies on manual efforts by experts, demanding a considerable amount of time and energy. This paper introduces a multi-scale feature fusion Siamese network guided by edge contour (MFS-GC) model. Constructing a Siamese network framework, it first uses a residual network to extract features of bone sticks, which is followed by computing the L2 distance for similarity measurement. During the extraction of feature vectors using the residual network, the BN layer tends to lose contour detail information, resulting in less conspicuous feature extraction, especially along fractured edges. To address this issue, the Spatially Adaptive DEnormalization (SPADE) model is employed to guide the normalization of contour images of bone sticks. This ensures that the network can learn multi-scale boundary contour features at each layer. Finally, the extracted multi-scale fused features undergo similarity measurement for local matching of bone stick fragment images. Additionally, a Conjugable Bone Stick Dataset (CBSD) is constructed. In the experimental validation phase, the MFS-GC algorithm is compared with classical similarity calculation methods in terms of precision, recall, and miss detection rate. The experiments demonstrate that the MFS-GC algorithm achieves an average accuracy of 95.5% in the Top-15 on the CBSD. The findings of this research can contribute to solving the rejoining issues of bone sticks.

Research on textile-vamp punching based on punching point gradient phase and edge significance

A visual technology to identify the punching points of a textile vamp to replace manual punching by machinery is proposed in this paper. This could solve several problems relating to the manual punching of textile vamp, such as high manual-punching strength, low efficiency, and poor punching accuracy. Unsharp mask guided filtering was adopted, to enhance the details of the textile-vamp punching points, considering the edge point position, gradient phase, and edge significance of the punching points, in building the template matching similarity measure function. A partial Hausdorff distance was proposed to sum and average values, to improve the degree of matching of punching point shape defects. A local search area of punching points to improve identification efficiency and punching evaluation criteria was established to evaluate the punching effects. The results showed that the matching similarity of the complete boundary was above 0.9. Positioning accuracy was 0.43 mm in the x-direction, 0.38 mm in the y-direction, and repeat positioning accuracy was 0.09 mm. The center average relative error was 5.03% and the relative error of the radius was 7.81%. Identification timeliness increased with the rotation angle, template size, and the number of punching points. When the rotation angle was between –180° and 180° and the number of punching points was 24, identification timeliness was 830 ms, which met productivity requirements. Textile-vamp punching grades A to D qualified, however, grades E and F were unqualified for punching.

Anti‐occlusion person re‐identification via body topology information restoration and similarity evaluation

AbstractIn real‐world scenarios, pedestrian images often suffer from occlusion, where certain body features become invisible, making it challenging for existing methods to accurately identify pedestrians with the same ID. Traditional approaches typically focus on matching only the visible body parts, which can lead to misalignment when the occlusion patterns vary. To address this issue and alleviate misalignment in occluded pedestrian images, the authors propose a novel framework called body topology information generation and matching. The framework consists of two main modules: the body topology information generation module and the body topology information matching module. The body topology information generation module employs an adaptive detection mechanism and capsule generative adversarial network to restore a holistic pedestrian image while preserving the body topology information. The body topology information matching module leverages the restored holistic image from body topology information generation to overcome spatial misalignment and utilises cosine distance as the similarity measure for matching. By combining the body topology information generation and body topology information matching modules, the authors achieve consistency in the body topology information features of pedestrian images, ranging from restoration to retrieval. Extensive experiments are conducted on both holistic person re‐identification datasets (Market‐1501, DukeMTMC‐ReID) and occluded person re‐identification datasets (Occluded‐DukeMTMC, Occluded‐ReID). The results demonstrate the superior performance of the authors proposed model, and visualisations of the generation and matching modules are provided to illustrate their effectiveness. Furthermore, an ablation study is conducted to validate the contributions of the proposed framework.

Cross-social networks analysis: building me-edge centered BUNet dataset based on implicit bridge users

PurposeThe intensive blooming of social media, specifically social networks, pushed users to be integrated into more than one social network and therefore many new “cross-network” scenarios have emerged, including cross-social networks content posting and recommendation systems. For this reason, it is mightily a necessity to identify implicit bridge users across social networks, known as social network reconciliation problem, to deal with such scenarios.Design/methodology/approachWe propose the BUNet (Bridge Users for cross-social Networks analysis) dataset built on the basis of a feature-based approach for identifying implicit bridge users across two popular social networks: Facebook and Twitter. The proposed approach leverages various similarity measures for identity matching. The Jaccard index is selected as the similarity measure outperforming all the tested measures for computing the degree of similarity between friends’ sets of two accounts of the same real person on two different social networks. Using “cross-site” linking functionality, the dataset is enriched by explicit me-edges from other social media websites.FindingsUsing the proposed approach, 399,407 users are extracted from different social platforms including an important number of bridge users shared across those platforms. Experimental results demonstrate that the proposed approach achieves good performance on implicit bridge users’ detection.Originality/valueThis paper contributes to the current scarcity of literature regarding cross-social networks analysis by providing researchers with a huge dataset of bridge users shared between different types of social media platforms.

A Fingerprint-Based Verification Framework Using Harris and SURF Feature Detection Algorithms

Amongst all biometric-based personal authentication systems, a fingerprint that gives each person a unique identity is the most commonly used parameter for personal identification. In this paper, we present an automatic fingerprint-based authentication framework by means of fingerprint enhancement, feature extraction, and matching techniques. Initially, a variant of adaptive histogram equalization called CLAHE (contrast limited adaptive histogram equalization) along with a combination of FFT (fast Fourier transform), and Gabor filters are applied to enhance the contrast of fingerprint images. The fingerprint is then authenticated by picking a small amount of information from some local interest points called minutiae point features. These features are extracted from the thinned binary fingerprint image with a hybrid combination of Harris and SURF feature detectors to render significantly improved detection results. For fingerprint matching, the Euclidean distance between the corresponding Harris-SURF feature vectors of two feature points is used as a feature matching similarity measure of two fingerprint images. Moreover, an iterative algorithm called RANSAC (RANdom SAmple Consensus) is applied for fine matching and to automatically eliminate false matches and incorrect match points. Quantitative experimental results achieved on FVC2002 DB1 and FVC2000 DB1 public domain fingerprint databases demonstrate the good performance and feasibility of the proposed framework in terms of achieving average recognition rates of 95% and 92.5% for FVC2002 DB1 and FVC2000 DB1 databases, respectively.

An accurate and distraction-free vision-based structural displacement measurement method integrating Siamese network based tracker and correlation-based template matching

Vision-based displacement measurement receives increasing attention on non-contact bridge monitoring while it faces challenges in long-time field applications due to the presence of environmental variations. To overcome this issue, this study proposes a novel distraction-free displacement measurement approach by integrating deep learning-based Siamese tracker with correlation-based template matching. The Siamese tracker used applies deep feature representations and learned similarity measures for image matching and also considers adaptive template update with time. Since the estimated bounding boxes by the Siamese tracker have size changes within frame sequences, a correction step is added to remove the centroid drifts between the template and the predicted target regions using correlation-based template matching. The proposed method is validated first in an indoor test and then implemented in monitoring tests on a short-span footbridge and a long-span road bridge, demonstrating its potential to handle challenging scenarios including partial occlusion, illumination changes, background variations and shade effects.

Adversarial-Metric Learning for Audio-Visual Cross-Modal Matching

Audio-visual matching aims to learn the intrinsic correspondence between image and audio clip. Existing works mainly concentrate on learning discriminative features, while ignore the cross-modal heterogeneous issue between audio and visual modalities. To deal with this issue, we propose a novel Adversarial-Metric Learning (AML) model for audio-visual matching. AML aims to generate a modality-independent representation for each person in each modality via adversarial learning, while simultaneously learns a robust similarity measure for cross-modality matching via metric learning. By integrating the discriminative modality-independent representation and robust cross-modality metric learning into an end-to-end trainable deep network, AML can overcome the heterogeneous issue with promising performance for audio-visual matching. Experiments on the various audio-visual learning tasks, including audio-visual matching, audio-visual verification and audio-visual retrieval on benchmark dataset demonstrate the effectiveness of the proposed AML model. The implementation codes are available on <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/MLanHu/AML</uri> .

An Adaptive Multisensor Image Fusion Method Based on Monogenic Features

Each kind of sensor is designed to adapt to a specific environment and usage scope. In this paper, an adaptive multisensor image fusion method based on monogenic features is proposed. For the proposed method, the monogenic features are extracted from the source images including monogenic amplitude, monogenic phase and monogenic orientation features. The energy matching similarity measure is used to measure the monogenic amplitude feature, and the complex coefficient structure similarity measure is used to characterize the monogenic phase and the monogenic orientation feature. Analyses the monogenic features in different types of regions, different fusion strategies are adopted for different characteristic regions. By means of visible and infrared image fusion, SAR and infrared image fusion experiments, the proposed method can effectively preserve a large amount of information and significantly improve the visual quality of the fused image, which can be evaluated with objective indicators.

Colour and orientation of pixel based video retrieval using IHBM similarity measure

Content-based video retrieval (CBVR) is the most energetic and stimulating research area since the early twentieth century in the domain of multimedia technology and immense quantity of retrieval techniques are introduced frequently. However, majority of the existent CBVR systems do not always give accurate results for all kinds of video databases with different colour, shape and texture feature descriptors. Sometimes, the images or videos that look similar are not semantically similar. Consequently, the retrieval outcomes that are solely centred on low level feature extraction are chiefly unsatisfactory and also unpredictable. This unlocks a new era for the research community to deviate the existent methodologies to new paradigm or direction that there is something at the back of the visual features which require to be regarded for precise searching and also retrieval. A novel CBVR methodology is suggested here centred on the selection of edge gradient feature descriptors known as HOG (Histograms of Oriented Gradients). HOG computes the edge gradient of the whole image, determines the orientation of every pixel and generates the histograms. Formerly, these extracted relevant histograms are utilized to retrieve the pertinent video frame as of the video sequence database through Integrated Histogram Bin Matching (IHBM) similarity measure. The Experimental Result of the proposed approach showed that the number of relevant retrieved video data samples is higher when compared to the existing HI Based CBVR system. The F1-score value is also high which in turn infers that the proposed approach’s performance is better when matched other existing approaches.

A Compliance Evaluation of a Hybridized Semantic Similarity Measure for Ontology Matching

A Compliance Evaluation of a Hybridized Semantic Similarity Measure for Ontology Matching

Research on Intelligent Semantic Search Base on HowNet

The HDWiki is same as the famous Wikipedia but major in Chinese include vase investment of manual effort and judgment, we divided this structured information into three categories: concept, relationship and entity. This paper de- scribe a novel symbol oriented search that facilitate user to format unambiguous query, especially to format complex logi- cal query which need several query to get answer on traditional search but just once here, it is building a bridge between what user know and what user wished, between the intent of user query and the understanding of search engine. This fa- cilitation and understanding base on the ternary relationships what are everywhere in the HDWiki and HowNet powered semantic similarity measure for knowledge matching. We realize a logical semantic oriented search system and conducted a detailed user study with users, and found that this logical semantic oriented search and its underlying knowledge base provides significant advantages over conventional search. It offers assistance to almost every query especially complex logical queries, making the query entry more clear and efficient for user as well as for search engine, broadening the proc- essing scope of complex logical query in the search, improving the relevance of the searched documents, and increased the probability of getting what user wished.

Web service discovery among large service pools utilising semantic similarity and clustering

ABSTRACTWith the rapid development of electronic business, Web services have attracted much attention in recent years. Enterprises can combine individual Web services to provide new value-added services. An emerging challenge is the timely discovery of close matches to service requests among large service pools. In this study, we first define a new semantic similarity measure combining functional similarity and process similarity. We then present a service discovery mechanism that utilises the new semantic similarity measure for service matching. All the published Web services are pre-grouped into functional clusters prior to the matching process. For a user’s service request, the discovery mechanism first identifies matching services clusters and then identifies the best matching Web services within these matching clusters. Experimental results show that the proposed semantic discovery mechanism performs better than a conventional lexical similarity-based mechanism.

Similarity measures of full polarimetric SAR images fusion for improved SAR image matching

Abstract. China’s first airborne SAR mapping system (CASMSAR) developed by Chinese Academy of Surveying and Mapping can acquire high-resolution and full polarimetric (HH, HV, VH and VV) Synthetic aperture radar (SAR) data. It has the ability to acquire X-band full polarimetric SAR data at a resolution of 0.5m. However, the existence of speckles which is inherent in SAR imagery affects visual interpretation and image processing badly, and challenges the assumption that conjugate points appear similar to each other in matching processing. In addition, researches show that speckles are multiplicative speckles, and most similarity measures of SAR image matching are sensitive to them. Thus, matching outcomes of SAR images acquired by most similarity measures are not reliable and with bad accuracy. Meanwhile, every polarimetric SAR image has different backscattering information of objects from each other and four polarimetric SAR data contain most basic and a large amount of redundancy information to improve matching. Therefore, we introduced logarithmically transformation and a stereo matching similarity measure into airborne full polarimetric SAR imagery. Firstly, in order to transform the multiplicative speckles into additivity ones and weaken speckles' influence on similarity measure, logarithmically transformation have to be taken to all images. Secondly, to prevent performance degradation of similarity measure caused by speckles, measure must be free or insensitive of additivity speckles. Thus, we introduced a stereo matching similarity measure, called Normalized Cross-Correlation (NCC), into full polarimetric SAR image matching. Thirdly, to take advantage of multi-polarimetric data and preserve the best similarity measure value, four measure values calculated between left and right single polarimetric SAR images are fused as final measure value for matching. The method was tested for matching under CASMSAR data. The results showed that the method delivered an effective performance on experimental imagery and can be used for airborne SAR matching applications.

Domain similarity based orthology detection

BackgroundOrthologous protein detection software mostly uses pairwise comparisons of amino-acid sequences to assert whether two proteins are orthologous or not. Accordingly, when the number of sequences for comparison increases, the number of comparisons to compute grows in a quadratic order. A current challenge of bioinformatic research, especially when taking into account the increasing number of sequenced organisms available, is to make this ever-growing number of comparisons computationally feasible in a reasonable amount of time. We propose to speed up the detection of orthologous proteins by using strings of domains to characterize the proteins.ResultsWe present two new protein similarity measures, a cosine and a maximal weight matching score based on domain content similarity, and new software, named porthoDom. The qualities of the cosine and the maximal weight matching similarity measures are compared against curated datasets. The measures show that domain content similarities are able to correctly group proteins into their families. Accordingly, the cosine similarity measure is used inside porthoDom, the wrapper developed for proteinortho. porthoDom makes use of domain content similarity measures to group proteins together before searching for orthologs. By using domains instead of amino acid sequences, the reduction of the search space decreases the computational complexity of an all-against-all sequence comparison.ConclusionWe demonstrate that representing and comparing proteins as strings of discrete domains, i.e. as a concatenation of their unique identifiers, allows a drastic simplification of search space. porthoDom has the advantage of speeding up orthology detection while maintaining a degree of accuracy similar to proteinortho. The implementation of porthoDom is released using python and C++ languages and is available under the GNU GPL licence 3 at http://www.bornberglab.org/pages/porthoda.Electronic supplementary materialThe online version of this article (doi:10.1186/s12859-015-0570-8) contains supplementary material, which is available to authorized users.

A Robust Appearance Model and Similarity Measure for Image Matching

&lt;div class=""abs_img""&gt; &lt;img src=""[disp_template_path]/JRM/abst-image/00270002/02.jpg"" width=""400"" /&gt; CP3 histogram&lt;/div&gt; An ideal similarity measure for matching image should be discriminative, producing a conspicuous correlation peak and suppressing false local maxima. Image matching tasks in practice, however, often involves complex conditions, such as blurring and fluctuating illumination. These may cause the similarity measure to not be discriminative enough. We utilized a robust scene modeling method to model the appearance of an image and propose an associated similarity measure for image matching. The proposed method utilizes a spatio-temporal learning stage to select a group of supporting pixels for each target pixel, then builds a differential statistic model of them to describe the uniqueness of the spatial structure and to provide illumination invariance for robust matching. We utilized this method for image matching in several challenging environments. Experimental results show that the proposed similarity measure produces explicit correlation peaks to achieve robust image matching. &lt;/span&gt;

A Fuzzy Logic Approach for Stereo Matching Suited for Real-Time Processing

In this paper, a novel method that uses both area and feature based information as similarity measures for stereo matching is proposed. Area-based information is suited for non-homogeneous regions while feature information helps in homogeneous areas. In order to define a conjugate pair, a fuzzy logic approach that combines the similarity information is used. The proposed method preserves discontinuities while reducing matching errors in homogeneous regions. This proposal is suited for real-time processing using dedicated hardware. We demonstrate and discuss performance using synthetic stereo pairs.

OCRM: Optimal Cost Region Matching Similarity Measure for Region Based Image Retrieval

Content Based Image Retrieval (CBIR) has been the most significant area in the applications of Pattern recognition and Computer Vision for the last three decades. However, there are many open problems left unresolved. Among these, one of the current problems of CBIR is to obtain an effective Similarity Measure. The CBIR systems make use of Integrated Region Matching (IRM) to match segmented images which is computationally economic, but it is not a metric distance whereas systems that use Minimum Cost Region Matching (MiCRoM) as a similarity measure is a metric distance, but computationally expensive. In order to address the above problem, this paper has developed the Optimal Cost Region Matching (OCRM) similarity measure for region based image retrieval. The proposed OCRM uses the north-west corner rule of the Transportation problem that fulfills the monge property. The experiment carried out on 1000 color images taken from the Corel database that are compared with IRM, and MiCRoM similarity measures.

Appearance Model Based Moving Object Matching across Disjoint Camera Views

A new object tracking scheme for multi-camera surveillance with non-overlapping views is proposed in this paper. Brightness transfer function (BTF) is used to establish relative appearance correspondence between different views. Mixtures of probabilistic principal component analysis (MPPCA) is incooperated to learn the subspace of brightness transfer function with the concern to deal with multiple different brightness areas in a scene. The incremental major color spectrum histogram (IMCSH) is used as similarity measure for reliable matching. Experimental results with real world videos show the effectiveness of the proposed algorithm.

QUANTIFYING NEARNESS IN VISUAL SPACES

Cybernetic vision systems can be deployed in problem domains where the goal is to achieve results similar to those produced by humans. Fundamentally, these problems consist of evaluation of image content between sets of images. This article contrasts two theoretical frameworks for image comparison, namely, the semantic similarity approach used in the earth mover's distance (EMD) and the integrated region matching (IRM) similarity measure, with the tolerance nearness measure (tNM) based on near set theory. The contribution of this article is a comparison of the image similarity measures EMD, IRM, and tNM, as well as a signature-based approach to calculating the tolerance nearness measure.

Similarity measures for image matching despite occlusions in stereo vision

In the context of computer vision, matching can be done with similarity measures. This paper presents the classification of these measures into five families. In addition, 18 measures based on robust statistics, previously proposed in order to deal with the problem of occlusions, are studied and compared to the state of the art. A new evaluation protocol and new analyses are proposed and the results highlight the most efficient measures, first, near occlusions, the smooth median powered deviation, and second, near discontinuities, a non-parametric transform-based measure, CENSUS.