Similarity-based Matching Research Articles

Bridge the gap between simulated and real-world data in optical fiber mode decomposition for accuracy improvement: A deep learning-based co-learning framework with visual similarity-based matching

In the area of optical fiber mode decomposition (MD), data-driven deep-learning (DL) algorithms have achieved remarkable strides with unlimited theoretically synthetic data, but the differences between theoretical and real-world data in the DL context are rarely investigated. The two differences, namely the information bottleneck and the domain discrepancy, between the simulation and realistic domain remain significant obstacles that limit the practical deployment of DL-based MD approaches in real-world optical systems. The paper aims to mitigate their influence on the accuracy limitation and the accuracy degradation of DL-based MD. First, we present a DL-based co-learning framework for MD for the first time. With the co-learning process, the knowledge of the teacher model learned from rich input information sources (near-field, far-field, and phase-distribution optical images) is transferred to the student model with the only near-field input source, which breaks the information bottleneck by allowing the student model to learn more knowledge in a simulation context. Furthermore, to settle the domain discrepancy between ideal data and practical observation, we propose a visual similarity-based matching to assign real values of modal coefficients to the practical optical images whose ground truth values cannot be easily acquired, which enables the transfer learning from the synthetic data to the real-world data for accuracy enhancement. Simulated and experimental results show that the proposed co-learning framework with visual similarity-based matching has realized high precision and robustness in MD simulated and practical tasks. By bridging the gap between theoretical and practical data, our study offers novel perspectives on DL-based MD practice.

Self-supervised structural similarity-based convolutional neural network for cardiac diffusion tensor image denoising.

Diffusion tensor imaging (DTI) is a promising technique for non-invasively investigating the myocardial fiber structures of human heart. However, low signal-to-noise ratio (SNR) has been a major limit of cardiac DTI to prevent us from detecting myocardium structure accurately. Therefore, it is important to remove the effect of noise on diffusion weighted (DW) images. Although the conventional and deep learning-based denoising methods have shown the potential to deal with effectively the noise in DW images, most of them are redundant information dependent or require the noise-free images as golden standard. In addition, the existed DW image denoising methods often suffer from problems of over-smoothing. To address these issues, we propose a self-supervised learning model, structural similarity based convolutional neural network with edge-weighted loss (SSECNN), to remove the noise effectively in cardiac DTI. Considering that the DW images acquired along different diffusion directions have structural similarity, and the noise in these DW images is independent and identically distributed, the structural similarity-based matching algorithm is proposed to search for the most similar DW images. Such similar noisy DW image pairs are then used as the input and target of the denoising network SSECNN, which consists of several convolutional and residual blocks. Through the self-supervised training with these image pairs, the network can restore the clean DW images and retain the correlations between the denoised DW images along different directions. To avoid the over-smoothing problem, we design a novel edge-weighted loss which enables the network to adaptively adjust the loss weights with iterations and therefore to improve the detail preserve ability of the model. To verify the superiority of the proposed method, comparisons with state-of-the-art (SOTA) denoising methods are performed on both synthetic and real acquired DTI datasets. Experimental results show that SSECNN can effectively reduce the noise in the DW images while preserving detailed texture and edge information and therefore achieve better performance in DTI reconstruction. For synthetic dataset, compared to the SOTA method, the root mean square error (RMSE), peak signal-to-noise ratio (PSNR), and structure similarity index measure (SSIM) between the denoised DW images obtained with SSECNN and noise-free DW images are improved by 6.94%, 1.98%, and 0.76% respectively when the noise level is 10%. As for the acquired cardiac DTI dataset, the SSECNN method could significantly improve SNR and contrast to noise ratio (CNR) of cardiac DW images and achieve more regular helix angle (HA) and transverse angle (TA) maps. The ablation experimental results validate that using the structure similarity-based method to search the similar DW image pairs yield the smallest loss, and with the help of the edge-weighted loss, the denoised DW images and diffusion metric maps can preserve more details. The proposed SSECNN method can fully explore the similarity between the DW images along different diffusion directions. Using such similarity and an edge-weighted loss enable us to denoise cardiac DTI effectively in a self-supervised manner. Our method can overcome the redundancy information dependence and over-smoothing problem of the SOTA methods.

Content-based medical image retrieval system for lung diseases using deep CNNs.

Content-based image retrieval (CBIR) systems are designed to retrieve images that are relevant, based on detailed analysis of latent image characteristics, thus eliminating the dependency of natural language tags, text descriptions, or keywords associated with the images. A CBIR system maintains high-level image visuals in the form of feature vectors, which the retrieval engine leverages for similarity-based matching and ranking for a given query image. In this paper, a CBIR system is proposed for the retrieval of medical images (CBMIR) for enabling the early detection and classification of lung diseases based on lung X-ray images. The proposed CBMIR system is built on the predictive power of deep neural models for the identification and classification of disease-specific features using transfer learning based models trained on standard COVID-19 Chest X-ray image datasets. Experimental evaluation on the standard dataset revealed that the proposed approach achieved an improvement of 49.71% in terms of precision, averaging across various distance metrics. Also, an improvement of 26.55% was observed in the area under precision-recall curve (AUPRC) values across all subclasses.

An interpretable natural language processing system for written medical examination assessment.

The assessment of written medical examinations is a tedious and expensive process, requiring significant amounts of time from medical experts. Our objective was to develop a natural language processing (NLP) system that can expedite the assessment of unstructured answers in medical examinations by automatically identifying relevant concepts in the examinee responses. Our NLP system, Intelligent Clinical Text Evaluator (INCITE), is semi-supervised in nature. Learning from a limited set of fully annotated examples, it sequentially applies a series of customized text comparison and similarity functions to determine if a text span represents an entry in a given reference standard. Combinations of fuzzy matching and set intersection-based methods capture inexact matches and also fragmented concepts. Customizable, dynamic similarity-based matching thresholds allow the system to be tailored for examinee responses of different lengths. INCITE achieved an average F1-score of 0.89 (precision = 0.87, recall = 0.91) against human annotations over held-out evaluation data. Fuzzy text matching, dynamic thresholding and the incorporation of supervision using annotated data resulted in the biggest jumps in performances. Long and non-standard expressions are difficult for INCITE to detect, but the problem is mitigated by the use of dynamic thresholding (i.e., varying the similarity threshold for a text span to be considered a match). Annotation variations within exams and disagreements between annotators were the primary causes for false positives. Small amounts of annotated data can significantly improve system performance. The high performance and interpretability of INCITE will likely significantly aid the assessment process and also help mitigate the impact of manual assessment inconsistencies.

Prophage Hunter: an integrative hunting tool for active prophages.

Identifying active prophages is critical for studying coevolution of phage and bacteria, investigating phage physiology and biochemistry, and engineering designer phages for diverse applications. We present Prophage Hunter, a tool aimed at hunting for active prophages from whole genome assembly of bacteria. Combining sequence similarity-based matching and genetic features-based machine learning classification, we developed a novel scoring system that exhibits higher accuracy than current tools in predicting active prophages on the validation datasets. The option of skipping similarity matching is also available so that there's higher chance for novel phages to be discovered. Prophage Hunter provides a one-stop web service to extract prophage genomes from bacterial genomes, evaluate the activity of the prophages, identify phylogenetically related phages, and annotate the function of phage proteins. Prophage Hunter is freely available at https://pro-hunter.bgi.com/.

Junction-Based Correspondence Estimation of Plant Point Cloud Data Using Subgraph Matching

Laser scanner-captured 3-D point cloud data analysis is becoming more commonly used for remote sensing and plant science applications. Because of nonrigidity and complexity, reconstructing a 3-D model of a plant is extremely challenging. Existing algorithms often fail to find correct correspondences for plantlike thin structures. We address the problem of finding 3-D junction points in plant point cloud data as a first step of this correspondence matching process. Temporarily, we transform the 3-D problem into 2-D by performing appropriate coordinate transformations to the neighborhood of each 3-D point. Our proposed method has two steps. First, a statistical dip test of multimodality is performed to detect the nonlinearity of the local 2D structure. Then, each branch is approximated by sequential random-sample-consensus line fitting and a Euclidean clustering technique. The straight line parameters of each branch are extracted using total-least-squares estimation. Finally, the straight line equations are solved to determine if they intersect in the local neighborhood. Such junction points are good candidates for subsequent correspondence algorithms. Using these detected junction points, we formulate a correspondence algorithm as a subgraph matching problem and show that, without using traditional descriptor similarity-based matching, good correspondences can be obtained by simply considering geodesic distances among graph nodes. Experiments on synthetic and real ( Arabidopsis plant) data show that the proposed method outperforms the state of the art.

Structural similarity‐based video fingerprinting for video copy detection

The authors propose a video fingerprinting method based on structural similarity and a graph model. Structural similarity-based fingerprint generation and double-layer matching are the two main contributions. The video is mapped to a graph with frames as its vertices, and structural similarity is proposed to compute the weights of the edges. Then, the fingerprint consisting of a match tag (a coarse fingerprint) and a fine fingerprint is generated by this graph. The match tag is generated by an independent set of this graph, and the fine fingerprint is generated by the weight matrix of the graph based on the two-block-dimensional discrete cosine transform. During the matching, the video can be matched at the first-layer using the match tag to obtain a candidate set, whereas the second-layer matching is performed in this candidate set using the fine fingerprint to find a final match. The proposed video fingerprinting method is shown to be resistant to geometric attacks on frames and impairment of transmission channels.

Content-based Image Retrieval (CBIR) using Hybrid Technique

retrieval is used in searching for images from images database. In this paper, content - based image retrieval (CBIR) using four feature extraction techniques has been achieved. The four techniques are colored histogram features technique, properties features technique, gray level co- occurrence matrix (GLCM) statistical features technique and hybrid technique. The features are extracted from the data base images and query (test) images in order to find the similarity measure. The similarity-based matching is very important in CBIR, so, three types of similarity measure are used, normalized Mahalanobis distance, Euclidean distance and Manhattan distance. A comparison between them has been implemented. From the results, it is concluded that, for the database images used in this work, the CBIR using hybrid technique is better for image retrieval because it has a higher match performance (100%) for each type of similarity measure so; it is the best one for image retrieval.

Cognitive architectures combine formal and heuristic approaches

Quantum probability (QP) theory provides an alternative account of empirical phenomena in decision making that classical probability (CP) theory cannot explain. Cognitive architectures combine probabilistic mechanisms with symbolic knowledge-based representations (e.g., heuristics) to address effects that motivate QP. They provide simple and natural explanations of these phenomena based on general cognitive processes such as memory retrieval, similarity-based partial matching, and associative learning.

Building linked ontologies with high precision using subclass mapping discovery

The creation of links between schemas of published datasets is a key part of the Linked Open Data (LOD) paradigm. The ability to discover these links "on the go" requires that ontology matching techniques achieve good precision and recall within acceptable execution times. In this paper, we add similarity-based and mediator-based ontology matching methods to the Agreementmaker ontology matching system, which aim to efficiently discover high precision subclass mappings between LOD ontologies. Similarity-based matching methods discover subclass mappings by extrapolating them from a set of high quality equivalence mappings and from the interpretation of compound concept names. Mediator-based matching methods discover subclass mappings by comparing polysemic lexical annotations of ontology concepts and by considering external web ontologies. Experiments show that when compared with a leading LOD approach, Agreementmaker achieves considerably higher precision and F-measure, at the cost of a slight decrease in recall.

WSMO-MX: A hybrid Semantic Web service matchmaker

The WSMO-MX service matchmaker applies different matching filters to retrieve Semantic Web services written in a dialect of the prominent service description language WSML-Rule. For this purpose, WSMO-MX recursively computes logic-based and syntactic similarity-based matching degrees and returns a ranked set of services that are semantically relevant to a given query. The matching filters perform ontology-based type matching, logical constraint matching, and syntactic matching. In this paper, we present the service description language WSML-MX, the hybrid matchmaker WSMO-MX for WSML services converted to WSML-MX, and the results of our experimental evaluation of its performance in terms of recall and precision over the test collection WSML-TC2.

Using evolution strategy for cooperative focused crawling on semantic web

Conventional focused crawling systems have difficulties on contextual information retrieval in semantic web environment. In order to deal with these problems, we propose a cooperative crawler platform based on evolution strategy to build semantic structure (i.e., local ontologies) of web spaces. Mainly, multiple crawlers can discover semantic instances (i.e., ontology fragments) from annotated resources in a web space, and a centralized meta-crawler can carry out incremental aggregation of the semantic instances sent by the multiple crawlers. To do this, we exploit similarity-based ontology matching algorithm for computing semantic fitness of a population, i.e., summation of all possible semantic similarities between the semantic instances. As a result, we could efficiently obtain the best mapping condition (i.e., maximizing the semantic fitness) of the estimated semantic structures. We have shown two significant contributions of this paper; (1) reconciling semantic conflicts between multiple crawlers, and (2) adapting to evolving semantic structures of web spaces over time.

Rule-based and case-based reasoning approach for internal audit of bank

Banks currently have a great interest in internal audits to reduce risks, to prevent themselves from insolvency, and to take quick action for financial incidents. This study presents an integrated audit approach of rule-based and case-based reasoning, which includes two stages of reasoning, i.e., screening stage based on rule-based reasoning and auditing stage based on case-based reasoning. Rule-based reasoning uses induction rules to determine whether a new problem should be inspected further or not. Case-based reasoning performs similarity-based matching to find the most similar case in case base to the new problem. The method presented is applied to data of internal audits of a bank.

A molecular-field-based similarity study of non-nucleoside HIV-1 reverse transcriptase inhibitors.

This article describes a molecular-field-based similarity method for aligning molecules by matching their steric and electrostatic fields and an application of the method to the alignment of three structurally diverse non-nucleoside HIV-1 reverse transcriptase inhibitors. A brief description of the method, as implemented in the program MIMIC, is presented, including a discussion of pairwise and multi-molecule similarity-based matching. The application provides an example that illustrates how relative binding orientations of molecules can be determined in the absence of detailed structural information on their target protein. In the particular system studied here, availability of the X-ray crystal structures of the respective ligand-protein complexes provides a means for constructing an 'experimental model' of the relative binding orientations of the three inhibitors. The experimental model is derived by using MIMIC to align the steric fields of the three protein P66 subunit main chains, producing an overlay with a 1.41 A average rms distance between the corresponding C alpha's in the three chains. The inter-chain residue similarities for the backbone structures show that the main-chain conformations are conserved in the region of the inhibitor-binding site, with the major deviations located primarily in the 'finger' and RNase H regions. The resulting inhibitor structure overlay provides an experimental-based model that can be used to evaluate the quality of the direct a priori inhibitor alignment obtained using MIMIC. It is found that the 'best' pairwise alignments do not always correspond to the experimental model alignments. Therefore, simply combining the best pairwise alignments will not necessarily produce the optimal multi-molecule alignment. However, the best simultaneous three-molecule alignment was found to reproduce the experimental inhibitor alignment model. A pairwise consistency index has been derived which gauges the quality of combining the pairwise alignments and aids in efficiently forming the optimal multi-molecule alignment analysis. Two post-alignment procedures are described that provide information on feature-based and field-based pharmacophoric patterns. The former corresponds to traditional pharmacophore models and is derived from the contribution of individual atoms to the total similarity. The latter is based on molecular regions rather than atoms and is constructed by computing the percent contribution to the similarity of individual points in a regular lattice surrounding the molecules, which when contoured and colored visually depict regions of highly conserved similarity. A discussion of how the information provided by each of the procedures is useful in drug design is also presented.