Local Matching Research Articles

Sending a Signal: Evaluating the Impact of Program and Geographic Preference Signaling on the Internal Medicine Residency Match.

Background For over a decade, the number of residency applications has surged, a trend known as "application inflation." COVID-19 further intensified this trend, leading the Association of American Medical Colleges (AAMC) to addressthe issue by introducing a supplemental application in the 2021-2022 cycle, allowing programs to identify applicants with a connection to their program or geographic region. For the 2022-2023 cycle, the number of program signals increased from five to seven. The impact of the supplemental application and the increase in signals on the likelihood of an applicant matching with a programhas yet to be evaluated. Methods This retrospective cohort study evaluated the impact of program signaling and geographic preference on the matching likelihood in our internal medicine residency program. Data from MyERAS® and the Supplemental Application for 640 applicants who applied to our large, urban, university-based program in the Southeastern United States during the 2020-2021 and 2022-2023 application cycles were included. Using univariate and multivariate analysis, we examined the correlation between program signal, geographic preference, and final match location. Results Applicants who sent a program signal had nearly three-foldhigher odds of matching with our program. Geographic preference was numerically but not statistically associated with higher odds of matching. Both signaling a preference for matching with a program in an urban environment and couples matching correlated with decreased odds of matching with our program. Geography was an important predictor of match location as residing in our AAMC geographic region, our four-state area, and our specific state had increased odds of matching with our program. Conclusions Signaling our program was associated with increased odds of matching with our program. Geographic preferences were less predictive of a match with our program; however, they did predict the likelihood of a match at a program within that region. Future studies are needed to ensure external validity.

IGU-Aug: Information-guided unsupervised augmentation and pixel-wise contrastive learning for medical image analysis.

Contrastive learning (CL) is a form of self-supervised learning and has been widely used for various tasks. Different from widely studied instance-level contrastive learning, pixel-wise contrastive learning mainly helps with pixel-wise dense prediction tasks. The counter-part to an instance in instance-level CL is a pixel, along with its neighboring context, in pixel-wise CL. Aiming to build better feature representation, there is a vast literature about designing instance augmentation strategies for instance-level CL; but there is little similar work on pixel augmentation for pixel-wise CL with a pixel granularity. In this paper, we attempt to bridge this gap. We first classify a pixel into three categories, namely low-, medium-, and high-informative, based on the information quantity the pixel contains. We then adaptively design separate augmentation strategies for each category in terms of augmentation intensity and sampling ratio. Extensive experiments validate that our information-guided pixel augmentation strategy succeeds in encoding more discriminative representations and surpassing other competitive approaches in unsupervised local feature matching. Furthermore, our pretrained model improves the performance of both one-shot and fully supervised models. To the best of our knowledge, we are the first to propose a pixel augmentation method with a pixel granularity for enhancing unsupervised pixel-wise contrastive learning. Code is available at https: //github.com/Curli-quan/IGU-Aug.

A novel deformation measurement method for rotating blade based on PSO-ILS image correlation matching and mismatch correction

Abstract Aiming at the problem of small measurement range and difficult measurement of traditional contact sensor method in the rotating state of aero-engine blades, this paper proposes a novel deformation measurement method for rotating blade based on image correlation matching and mismatch correction. Firstly, a Particle Swarm Optimization-Iterative Local Search image intelligent matching algorithm is proposed, which effectively balances global search and local optimization, and the image matching displacement accuracy reaches 10−3 pixel. Secondly, a mismatch point detection method based on multi-scale local Root Mean Square is proposed, and the high-precision detection of mismatch points is realized by considering the influence of local structure. Finally, through the hierarchical refinement of the sub-pixel level mismatch point correction method, the mismatch point is corrected in the two iterative stages of global search and local optimization. The corrected mismatch point response value K is reduced by 99% compared with that before correction, which further improves the accuracy of deformation field calculation. In the experiment, the deformation of the rotating blade of 4500 RPM is measured, which proves that the image intelligent matching algorithm and the deformation field calculation method proposed in this paper can provide new methods and technical support for the accurate measurement of the blade surface deformation.

Graphite: painting genomes using a colored de Bruijn graph.

The recent growth of microbial sequence data allows comparisons at unprecedented scales, enabling the tracking of strains, mobile genetic elements, or genes. Querying a genome against a large reference database can easily yield thousands of matches that are tedious to interpret and pose computational challenges. We developed Graphite that uses a colored de Bruijn graph (cDBG) to paint query genomes, selecting the local best matches along the full query length. By focusing on the best genomic match of each query region, Graphite reduces the number of matches while providing the most promising leads for sequence tracking or genomic forensics. When applied to hundreds of Campylobacter genomes we found extensive gene sharing, including a previously undetected C. coli plasmid that matched a C. jejuni chromosome. Together, genome painting using cDBGs as enabled by Graphite, can reveal new biological phenomena by mitigating computational hurdles.

LightSGM: Local feature matching with lightweight seeded

Addressing the quintessential challenge of local feature matching in computer vision, this study introduces a novel fast sparse seed graph structure named LightSGM. This structure aims to refine the characterization of graph features while minimizing superfluous connections. Initially, a subset of high-quality seed feature points is curated using a confidence filter. Subsequently, keypoint features are assimilated into this seed set via graph pooling, and the composite features are further processed through a memory and computation-efficient seed transformer to capture rich contextual information about the keypoints. The seed feature points are then relayed back to the original keypoints using an inverse process known as graph unpooling. The paper also introduce an adaptive mechanism to determine the optimal number of model layers based on the intricacy of matching image pairs. A Matching Point Prediction Header is employed to extract the final set of matching points. Through extensive experimentation on image matching and position estimation, LightSGM has demonstrated its prowess in delivering competitive matching accuracy while maintaining a balance with real-time processing capabilities.

Tracing participants for longitudinal environmental health research using social networking sites: a pilot study

Aim: Longitudinal cohort study designs are considered the gold standard for investigating associations between environmental exposures and human health yet they are characterized by limitations including participant attrition, and the resource implications associated with cohort recruitment and follow-up. Attrition compromises the integrity of research by threatening both the internal and external validity of empirical results, weakening the accuracy of statistical inferences and the generalizability of findings. This pilot study aimed to trace participants from a historical cohort study, the Hamilton Child Cohort Study (HCC) (n = 3,202), (1976–1986, 2003–2008) which was originally designed to examine the relative contribution of indoor and outdoor exposure to air pollution on respiratory health. Methods: Original participants were traced through social networking sites (SNS) by leveraging personal identifying data (name, age, sex, educational affiliation, and geographical locations) from the HCC entered into SNS search engines. Results: Of the original cohort (n = 3,166), 21% (n = 665) were identified as having social media presence (SMP) on a single social media platform, with 15% (n = 479) found on Facebook, 6% (n = 185) on LinkedIn, < 1% (n = 9) on Instagram, and n = 1 participant on Twitter. However, 68% (n = 2,168) of the cohort were associated with multiple SNS with the same features (matching names, ages, and locations), making conclusive identification challenging. The remaining 11% (n = 334) of the cohort had no SMP (NSMP). Statistical differences in sample characteristics of each cohort were analyzed using the Pearson chi-square test. Significant differences between the SMP and NSMP cohorts were found in relation to sex (p < 0.001), and childhood neighborhood of residence (p < 0.05). Conclusions: This study underscores social media’s potential for tracing participants in longitudinal studies while advising a multi-faceted approach to overcome inherent limitations and biases. A full-scale study is necessary to determine whether utilizing SNS to trace participants for longitudinal research is an effective tool for re-engaging research participants lost to attrition.

A Self-Distillation Contrastive Learning Architecture for Global and Local Underwater Terrain Feature Extraction and Matching

A Self-Distillation Contrastive Learning Architecture for Global and Local Underwater Terrain Feature Extraction and Matching

Development and Evaluation of a Two-Staged 3D Keypoint Based Workflow for the Co-Registration of Unstructured Multi-Temporal and Multi-Modal 3D Point Clouds

Abstract. Robust and automated point cloud registration methods are required in many geoscience applications using multi-temporal and multi-modal 3D point clouds. Therefore, a 3D keypoint-based coarse registration workflow has been implemented, utilizing the ISS keypoint detector and 3DSmoothNet descriptor. This paper contributes to keypoint-based registration research through variations of the standard workflow proposed in the literature, applying a two-staged strategy of global and local keypoint matching as well as prototypical keypoint projection and fine registration based on ICP. Further, by testing the utilized detector and descriptor on unstructured, multi-temporal and multi-source point clouds with variations in point cloud density, generalization ability is tested outside benchmark data. Therefore, data of the Bøverbreen glacier in Jotunheimen, Norway has been acquired in 2022 and 2023, deploying UAV-based image matching and terrestrial laser scanning. The results show good performance of the implemented robust matching algorithm PROSAC, requiring fewer iterations than the well-known RANSAC approach, but solving the rigid body transformation with TEASER++ is faster and more robust to outliers without demanding pre-knowledge of the data. Further, the results identify the keypoint detection as most limiting factor in speed and accuracy. Summarizing, keypoint-based coarse registration on low density point clouds, applying a global and local matching strategy and transformation estimation using TEASER++ is recommended. Keypoint projection shows potential, increasing number and precision in low density clouds, but has to be more robust. Further research needs to be carried out, focusing on identifying a fast and robust keypoint detector.

Reference-based super-resolution reconstruction of remote sensing images based on a coarse-to-fine feature matching transformer

Remote sensing image super-resolution reconstruction technology mines the deep details of remote sensing image, which has been widely used in the fields of intelligent and precise agriculture, intelligent transportation, earth surface object recognition, and so on. To obtain more detailed information,we design a reference-based super-resolution reconstruction network. Firstly, the coarse-to-fine feature matching strategy is adopted for the features of the input image. Global coarse matching is performed on the center patch of each block, and then pixel-level local fine matching is performed on the edge patches of the block. This both reduces the amount of computation and improves the matching accuracy. A threshold is set to determine whether the feature matching results meet the criteria for feature transfer. Finally, different scale features are fused through several convolutional layers and sampling operations, obtaining the reconstruction features after a fourfold increase in resolution. The ultimate super-resolution image is generated through a decoder. We have performed training and testing on remote sensing datasets. Compared to the current state-of-the-art methods, our proposed method is visually more details and outperforms other methods in terms of objective evaluation metrics.

CRTrack: Learning Correlation-Refine network for visual object tracking

Conducting reliable feature interaction plays a critical role in the visual tracking community, especially in recent dominated Siamese-based tracking paradigm. In general, there are two primary approaches for fusing representations from template and search area in the Siamese setting, i.e., cross-correlation and transformer modeling. The former provides a straightforward interaction solution, which may have limitations in handling complex scenarios, such as appearance variations and occlusion. While the latter offers an effective interaction mechanism, albeit with higher computation complexity and model cost. In contrast to traditional Siamese-based trackers which rely on two mentioned feature cross-correlation operators, this paper proposes a novel Correlation-Refine network to address the issue of lacking semantic information caused by local linear matching in correlation, from both spatial and channel perspectives. Correlation-Refine network (named CR) is solely built on top of fully convolutional layers, without employing intricate transformer mechanisms or complex methods to fuse features from multiple scales. Moreover, we present a concise yet effective convolutional tracking framework based on the correlation-refine network. CR network can increase the discriminative ability of semantic information in a coarse-to-fine manner: it gradually learns the semantic features of the target to be tracked and suppresses interference from similar objects by stacking multiple CR layers. Extensive experiments and comparisons with recent competitive trackers in challenging large-scale benchmarks demonstrate that, our tracker outperforms all previous convolutional trackers and has competitive results with transformer-based method. The code will be made available.

Novel Illumination-invariant Face Recognition Approach via Reflectance-luminance and Local Matching Model with Weighted Voting System

In this study, a novel approach has been introduced for face recognition that is unaffected by changes in illumination. This method is based on the reflectance-luminance model and incorporates local matching using a weighted voting technique to eliminate any artifacts present in the retina images. A total of 37 different linear and nonlinear filters were tested, including high pass and low pass filters, to achieve the reflectance component of the image, which remains invariant to changes in illumination. Among these filters, the maximum filter, which is a simple filter with low computational complexity, yielded the best results in extracting the illumination invariants. It was observed that the illumination invariants obtained through this method outperformed other methods such as QI, SQI, and image enhancement techniques in terms of recognition accuracy. Importantly, the proposed method does not require any prior knowledge about the facial shape or illumination conditions and can be applied to individual images independently. Unlike many existing methods, this approach does not rely on multiple images during the training stage and does not require any parameter selection to generate the illumination invariants. To further enhance the robustness of illumination, a weighted voting system was introduced. Certain regions of the image, which may adversely affect the recognition outcome due to poor illumination, occlusion, noise, or lack of distinctive information, were identified using predefined factors such as grayscale mean, image entropy, and mutual information. The proposed method was also compared to other face recognition methods in the presence of occlusions, and it demonstrated promising results outperforming existing methods. The Python algorithm successfully detects obstructed faces, genders, and ages in videos with a face matching accuracy between 80.9% to 96.9% based on proximity.

Direct Implantation of Patient Brain Tumor Cells into Matching Locations in Mouse Brains for Patient-Derived Orthotopic Xenograft Model Development.

Background: Despite multimodality therapies, the prognosis of patients with malignant brain tumors remains extremely poor. One of the major obstacles that hinders development of effective therapies is the limited availability of clinically relevant and biologically accurate (CRBA) mouse models. Methods: We have developed a freehand surgical technique that allows for rapid and safe injection of fresh human brain tumor specimens directly into the matching locations (cerebrum, cerebellum, or brainstem) in the brains of SCID mice. Results: Using this technique, we successfully developed 188 PDOX models from 408 brain tumor patient samples (both high-and low-grade) with a success rate of 72.3% in high-grade glioma, 64.2% in medulloblastoma, 50% in ATRT, 33.8% in ependymoma, and 11.6% in low-grade gliomas. Detailed characterization confirmed their replication of the histopathological and genetic abnormalities of the original patient tumors. Conclusions: The protocol is easy to follow, without a sterotactic frame, in order to generate large cohorts of tumor-bearing mice to meet the needs of biological studies and preclinical drug testing.

The Effect of Situational Variables on Women’s Rink Hockey Match Outcomes

The main objective of the present study was to develop a concise predictive model to determine the likelihood of winning in female rink hockey based on various situational variables. Additionally, the study aimed to assess the individual impact of each predictor on match outcomes. The analysis encompassed a dataset of 840 matches during five consecutive seasons (from 2018–2019 to 2022–2023) in the Spanish first division (OkLiga). Employing the comprehensive method of all possible regressions, the most effective predictive logistic model for match outcomes was identified. This entire model featured five categorical predictor variables (match location, team level, opponent level, scoring first, and match status at halftime) and one binary outcome variable (match outcome). Subsequently, the final model, which exhibited a sensitivity and specificity surpassing 80% for a cut-off point of 0.439, emerged. This model was applied to predict winning a match in 18 frequent situations determined from a two-step cluster analysis. Within this predictive framework, match status at halftime emerged as the most influential predictor impacting the match outcome, followed by opponent level, team level, and match location. The implications of our findings extend to rink hockey coaches and practitioners. Recognizing the significant impact of situational variables on match outcomes empowers them to customize game plans and design more specific strategies, thereby enhancing game understanding and elevating the overall performance.

Enhancing Regular Expression Processing through Field-Programmable Gate Array-Based Multi-Character Non-Deterministic Finite Automata

This work investigates the advantages of FPGA-based Multi-Character Non-Deterministic Finite Automata (MC-NFA) for enhancing regular expression processing over traditional software-based methods. By integrating Field-Programmable Gate Arrays (FPGAs) within a data processing framework, our study showcases significant improvements in processing efficiency, accuracy, and resource utilization for complex pattern matching tasks. We present a novel approach that not only accelerates database and network security applications, but also contributes to the evolving landscape of computational efficiency and hardware acceleration. The findings illustrate that FPGA’s coherent access to main memory and the efficient use of resources lead to considerable gains in processing times and throughput for handling regular expressions, unaffected by expression complexity and driven primarily by dataset size and match location. Our research further introduces a phase shift compensation technique that elevates match accuracy to optimal levels, highlighting FPGA’s potential for real-time, accurate data processing. The study confirms that the benefits of using FPGA for these tasks do not linearly correlate with an increase in resource consumption, underscoring the technology’s efficiency. This paper not only solidifies the case for adopting FPGA technology in complex data processing tasks, but also lays the groundwork for future explorations into optimizing hardware accelerators for broader applications.

SADGFeat: Learning local features with layer spatial attention and domain generalization

Local feature plays a pivotal role in various robotic tasks, including 3D reconstruction and visual localization. Although deep learning-based local features have proven superior to their traditional counterparts, they still face challenges in practical applications due to matching failures. These challenges primarily stem from inaccuracies in keypoint localization and the limited robustness of descriptors particularly with significant appearance changes. In this study, we introduce a novel method for local feature learning based on layer spatial attention and domain generalization. Firstly, a keypoint extraction strategy driven by layer spatial attention from high-level feature is proposed to enhance the accuracy of keypoint localization, progressing from coarse to fine granularity. Secondly, a new learning paradigm based on domain generalization is developed to extract local features with resilience against variations in illumination. To enrich the domain diversity within the training dataset, a real-time and lossless Fourier transform-based domain augmentation method is introduced. This method seamlessly integrates into the training process, enhancing the model's adaptability to varying domains. Additionally, explicit feature alignment-based representation learning is performed, further reinforcing the extraction of domain-invariant local features. Experimental results on public datasets demonstrate that the proposed method achieves state-of-the-art performance across various downstream tasks reliant on local feature matching, such as image matching, 3D reconstruction, and long-term visual localization.

Are there meaningful changes in the performance of professional soccer teams from the first half of the match to the second? Evidence from the data of eight consecutive seasons of the Chinese Super League

ABSTRACT In the aim to better understanding the dynamic of performance changes between the first and second halves of professional soccer teams, statistics of 1899 matches from 2012 to 2019 of the Chinese Super League (CSL) were collected and analysed. The generalised mixed linear modelling was used to estimate the mean change of each of 40 physical and technical performance-related parameters of soccer teams from the first half of the match to the second, controlling the effects of match status, match location, team strength and opponent strength. Results showed that, out of the selected 40 parameters, only goals (first half vs second half M ± SD: 0.60 ± 0.76 vs 0.82 ± 0.92, ES; ±99% confidence limits: 0.21; ±0.05), shots (5.7 ± 2.8 vs 6.8 ± 3.3, 0.34; ±0.05), individual possession in the attacking third (55 ± 21 vs 62 ± 25, 0.31; ±0.05), cross (9.7 ± 4.6 vs 10.9 ± 5.8, 0.24; ±0.06) and yellow cards (0.75 ± 0.84 vs 1.24 ± 1.07, 0.45; ±0.06) showed substantially increment from the first half to the second, other parameters presented only trivial changes. Hence, it is concluded that the physical performance of soccer teams in the second half did not show meaningful change, while technical actions related to attacking tended to increase. The findings can offer soccer coaches valuable insights into when and how to adjust their tactics, manage their players, thus can maintain or enhance the team’s match performance.

EMD AND WAVELET-BASED ECG SIGNAL DENOISING AND QRS COMPLEX DETECTION

This project presents a novel method for noise removal and QRS complex detection in electrocardiogram (ECG) signals, employing Empirical Mode Decomposition (EMD), windowing, and wavelet thresholding techniques. Noise is added to the ECG signal, and EMD is applied to obtain denoised components. Local minima detection and peak matching identify QRS complex boundaries. Windowing isolates the QRS regions, followed by wavelet thresholding for further denoising. Evaluation metrics like improved Signal-to-Noise Ratio, Mean Square Error, and Percent Root Mean Square Difference are calculated at different input SNR values, demonstrating the method's effectiveness in preserving the QRS complex while removing noise.. Keywords: ECG signal processing, QRS complex detection, Empirical Mode Decomposition (EMD), Windowing, Wavelet thresholding, QRS complex extraction, Signal-to-Noise Ratio (SNR), Mean Square Error (MSE), Percent Root Mean Square Difference (PRD).

A Hierarchical Utilization of Semantic Gradients and Scene Structure for Visual Place Recognition

Visual Place Recognition (VPR) is a fundamental element for long-term Simultaneous Localization and Mapping (SLAM) systems. For long-term VPR, severe appearance and viewpoint variations are inevitable. In this paper, we introduce a novel VPR system named Semantic Scene Structure Place Recognition (3SPR), inspired by the repeatability of semantic gradients and the scene structure of urban environments. Semantic gradients are densely sampled according to the sum of absolute gradients of all channels in the logits layer. Features of the semantic gradients in different layers are concatenated to exploit features’ characteristics at different levels. Based on partitions by vanishing points of road lines and Vector of Locally Aggregated Descriptors (VLAD), the Scene Structure VLAD (SSVLAD) is generated from concatenated features of the semantic gradients. Moreover, a local point group match method is used to enhance the spatial verification. Experimental results show that our method achieves state-of-the-art performance on the Oxford Robotcar dataset and the Synthia dataset.

Influence of contextual factors on match running performance of starters and non-starters in elite youth male soccer players.

This study investigated the effects of match location, quality of opposition (classified into: strong [1st to 4th position]; intermediate [5th to 15th position]; weak [16th to 20th position]), and match outcome on the match running performance of starters and non-starters from a top elite Brazilian soccer team. Absolute measures were calculated using total distance, high-speed (19.8-25.2 km·h-1), sprinting (≥ 25.2 km·h-1), total distance high-acceleration (> 2m·s2), and deceleration (< -2m·s2) were recorded by GPS units from a sample of young soccer players (N = 25) in a total of 17 matches. Relative measures were calculated by dividing absolute measures by the total duration of the matches. Non-starters covered greater total distance (p = 0.02), sprinting (p = 0.02), high-acceleration and deceleration (p = 0.04), sprinting distance relative per minute played (p = 0.005), and high-acceleration and deceleration relative per minute played (p < 0.001) when the team plays at home, strong opponents, and wins the matches. Starters covered greater total distance in high-speed running (p = 0.04), high-acceleration and deceleration (p = 0.03), and high-speed running relative per minute played (p = 0.04) when the team plays strong opponents and wins the matches. These findings highlight the impact of contextual factors during matches on the locomotor performance of young soccer players.

Kernel based local matching network for video object segmentation

Kernel based local matching network for video object segmentation