Correspondence Search Research Articles

ShellNeRF: Learning a Controllable High‐resolution Model of the Eye and Periocular Region

AbstractEye gaze and expressions are crucial non‐verbal signals in face‐to‐face communication. Visual effects and telepresence demand significant improvements in personalized tracking, animation, and synthesis of the eye region to achieve true immersion. Morphable face models, in combination with coordinate‐based neural volumetric representations, show promise in solving the difficult problem of reconstructing intricate geometry (eyelashes) and synthesizing photorealistic appearance variations (wrinkles and specularities) of eye performances. We propose a novel hybrid representation ‐ ShellNeRF ‐ that builds a discretized volume around a 3DMM face mesh using concentric surfaces to model the deformable ‘periocular’ region. We define a canonical space using the UV layout of the shells that constrains the space of dense correspondence search. Combined with an explicit eyeball mesh for modeling corneal light‐transport, our model allows for animatable photorealistic 3D synthesis of the whole eye region. Using multi‐view video input, we demonstrate significant improvements over state‐of‐the‐art in expression re‐enactment and transfer for high‐resolution close‐up views of the eye region.

Design and Development of College Archives and Correspondence Applications

This study aims to develop a web-based archiving and correspondence application to improve the efficiency and effectiveness of archiving and delivery of correspondence on the smart campus. The method used is the software development method. This research involves data collection through interviews and observations to understand the process of archiving and delivering correspondence on smart campus. The developed application can help users to archive and deliver correspondence more efficiently and effectively with features such as correspondence search, automatic archiving, and task and notification settings. The results showed that the use of web-based archives and correspondence applications can increase the efficiency and effectiveness of archiving and delivering correspondence on smart campuses by reducing time and costs, and increasing the accuracy and security of correspondence and archive management.

An overlap estimation guided feature metric approach for real point cloud registration

Real point cloud registration, involving homologous and cross-source 3D data, poses significant challenges such as partial overlap, high noise, density disparities, and scale variations. In this paper, we introduce a framework called OEFM to guide transform estimation on the basis of feature metrics through overlap estimation with overlap region center weighting, which estimates transform parameters by minimizing feature projection error on overlap region features. Our proposed approach utilizes an overlap filtering module with a center-weighting mechanism to filter overlap points and feeds these points into a feature metric framework with a forward–backward transformation to estimate the transformation parameters. The rationale behind our approach is that only overlapping regions are useful for point cloud registration, while non-overlapping regions are anomalous. Furthermore, our proposed approach requires no correspondence search, making it robust to partial overlaps, large noise, density differences, and scale variations. We show that our approach achieves the highest registration recall for 3DMatch and 3DCSR in extensive experiments on both homologous and cross-source datasets.

Multipluralism with education in Google books: Mechanisms descriptive about transsexuality and social representations and rights of adolescents

The work presented was prepared according to the need to understand the representations of human beings in their recognized historical rights identified throughout mainly adolescence as an important stage of life. This research is also motivated to understand public health policies identified in alignment with the representations of the human being in the phases of life, especially those aimed at transsexuals. I used the following lexical items added in English using the search engine through Google Books and, concomitantly, scanning in Brazilian Portuguese in a correspondence search engine called Brazilian Corpus made available by the Linguateca1 initiative. Therefore, these are recognized masculine genders, women, gays, lesbians, homosexuals, transgender people; person, transgender, transgender person, transsexual person, spiritual person, transgender person, transsexual person, gender queer, gender dysphoria, gender identity, gender diverse, gender non-conforming minors, gender non-conforming, gender affirming, non-binary; non-binary; transgender, gender diversity, non-conforming adolescents, LGBTQ, non- conforming people, minority genders, gender transition, family, adopted family (as well as their own plural forms). The phases of life have been operationalized by the following measures: man, woman, adolescent, adolescence, adult, boy, girl, child, elderly, child, young person and adolescent (as well as their plural forms). The data for the analysis consists of publications made available by Google Books in the period 1800 to 2008, totaling approximately 500 million words, in addition to the use of the SKETCH ENGINE Corpus which brings together publications in English and Brazilian Portuguese in this analysis. The Portuguese Corpus contains around 1 billion words of data from newspapers and magazines on the Internet in four Portuguese-speaking countries between 2012 and 2019 and is called the Portuguese Web (ptTenTen20). To this end, several analyzes were carried out, focusing in particular on the results related to the rights of transsexual people and when these rights began to be recognized in publications and the possible health policies addressed to adolescents and education.

DIFFICULTIES IN TRANSLATING ARCHAISMS AND HISTORICISMS

The article examines archaisms and historicisms in the English language of W.Shakespeare’s time as a translation challenge, as well as the functional element of archaic vocabulary and methods of transmission when translating from English into Uzbek. Various theoretical viewpoints in the realm of archaism and historicism translation are addressed. Based on their research, the authors conclude that in order to appropriately express the traits of a distant literary text, it is critical to keep, first and foremost, the lexical and grammatical characteristics of the original while translating it. The key lexical characteristics of such a text are archaisms and historicisms. After researching the various methods of translating these lexical units from English intoUzbek, the authors believe that the selection of an analogue and the search for equivalent correspondences are the most effective.

Fast-SNARF: A Fast Deformer for Articulated Neural Fields.

Neural fields have revolutionized the area of 3D reconstruction and novel view synthesis of rigid scenes. A key challenge in making such methods applicable to articulated objects, such as the human body, is to model the deformation of 3D locations between the rest pose (a canonical space) and the deformed space. We propose a new articulation module for neural fields, Fast-SNARF, which finds accurate correspondences between canonical space and posed space via iterative root finding. Fast-SNARF is a drop-in replacement in functionality to our previous work, SNARF, while significantly improving its computational efficiency. We contribute several algorithmic and implementation improvements over SNARF, yielding a speed-up of 150×. These improvements include voxel-based correspondence search, pre-computing the linear blend skinning function, and an efficient software implementation with CUDA kernels. Fast-SNARF enables efficient and simultaneous optimization of shape and skinning weights given deformed observations without correspondences (e.g. 3D meshes). Because learning of deformation maps is a crucial component in many 3D human avatar methods and since Fast-SNARF provides a computationally efficient solution, we believe that this work represents a significant step towards the practical creation of 3D virtual humans.

A survey on conventional and learning‐based methods for multi‐view stereo

Abstract3D reconstruction of scenes using multiple images, relying on robust correspondence search and depth estimation, has been thoroughly studied for the two‐view and multi‐view scenarios in recent years. Multi‐view stereo (MVS) algorithms aim to generate a rich, dense 3D model of the scene in the form of a dense point cloud or a triangulated mesh. In a typical MVS pipeline, the robust estimations for the camera poses along with the sparse points obtained from structure from motion (SfM) are used as input. During this process, the depth of generally every pixel of the scene is to be calculated. Several methods, either conventional or, more recently, learning‐based have been developed for solving the correspondence search problem. A vast amount of research exists in the literature using local, global or semi‐global stereomatching approaches, with the PatchMatch algorithm being among the most popular and efficient conventional ones in the last decade. Yet, and despite the widespread evolution of the algorithms, yielding complete, accurate and aesthetically pleasing 3D representations of a scene remains an open issue in real‐world and large‐scale photogrammetric applications. This work aims to provide a concrete survey on the most widely used MVS methods, investigating underlying concepts and challenges. To this end, the theoretical background and relative literature are discussed for both conventional and learning‐based approaches, with a particular focus on close‐range 3D reconstruction applications.

Contributions to the Translation of Paremias

The unique nature of paremias or short, sententious sentences, presents several translation problems. This article shows how the use of ICTs to the study and preservation of paremias contributes to the existence of different types of consultation sources (journals, monographs, repertoires, didactic proposals), which facilitates both the search for correspondences as well as the teaching / learning of them in the translation classroom. After explaining the concept of the word “paremia” and the typology of the different types of paremias, we explain the translation techniques that have been disseminated thanks to ICTs, to afterwards present different publications on paremias freely available on the Internet: a selection of multilingual databases, articles in journals, activities and didactic units for the translation class. Particular emphasis will be placed on reference sources concerning paremias in Spanish, such as multilingual databases: the Gotzon Garate collection (Fundación Bilbao Bizkaia Kutxa), the Refranero multilingüe (Centro Virtual Cervantes, Cervantes Institute) and ParemioRom (University of Barcelona).

Antiquities of the Vidivarii

No attempts have been previously made to discover the antiquities of the Vidivarii. Their discovery is the purpose of this article. The relevance of this work is that the discovery of these antiquities will be the key to understanding the history of the southeastern Baltic on the threshold of the Middle Ages. The methodology for searching for correspondence between historical and archeological data has been developed in European archeology in identifying the regions of the Germanic tribes of the era of Roman influence. The research method was search for correspondence between the features of the archaeological material and the data of a written source that tells about the Vidivarii. First of all, it is necessary to consider the funerary monuments of the “island” of Gepedoyos and its environs in the first half of the 5th century AD. They are represented mainly by urnless cremations (fragments of calcined bones are scattered in the remains of a funeral pyre) under stonework. Artifacts that can be associated with the activity of the Vidivarii are massive hryvnias mainly with receding, expanding ends, coin and clothing hoards of the first half of the 5th century AD, three-beam brooches and plate clasps, biconical bowls and circular ceramics, knife-daggers. We draw a conclusion that the zone of settlement of German soldiers was initially located on the “island of Gepedoyos”, which is evidenced by the location of a large number of coin and clothing treasures on this hill. Archaeological material shows the connections of the Vidivarii both with the west of the Baltic and with the Middle Danube. There is an evidence of the origin of the Vidivarii from the array of Germanic tribes on the western coast of the Baltic and the way in which groups of Germanic warriors in the era of Attila fell to the ground in the delta of the river Vistula. Traces of activity of the Vidivarii in the archaeological material of the coast of the Vistula/Kaliningrad Bay are represented by finds of Late Roman multiplis, solidi, and gold hryvnias of the R300 type. The process of participation of the Vidivarii (in the form of a kind of catalyst) in the formation of Prussian archaeological culture at its early stage is represented by animalheaded brooches, clasps with a star-shaped stem, and other items characteristic of the late 5th-6th centuries only for the Witland region. Thus, discrete archaeological materials have been identified that are confidently associated with the Vidivarii.

Translations of the Qurʼān into Chinese: Peculiarities of the Names of Suras

The paper discusses some peculiarities of translating the names of suras of the Qurʼān into Chinese based on the study of the most common Chinese translation made in the second half of the last century by the Muslim intellectual Muhammad Ma Jian. The work examines suras possessing the most significant discrepancies between the original Arabic names and the corresponding Chinese equivalents. The study is accompanied by commentary on the potential reasons for these variations while taking into account factors of an ethnocultural and linguistic nature. In a number of cases, comparisons with translations of the Qurʼān into Russian are also provided. The interdisciplinary nature of the work is manifested in the application of methods of source study, historical and philological, and comparative analysis. Having researched the main milestones in the history of translations of the Qurʼān into Chinese in the context of the spread of Islam in China, the authors come to the conclusion that the peculiarities of the translation of the names of suras into Chinese are largely due to the history of their occurrence and the variability of the original names since the written record of the Holy Text. At the same time, Muhammad Ma Jian’s belonging to the Hui community and Chinese traditions and customs also influenced the process of choosing equivalents. Obviously, the search for correspondences relied on the synthesis of methods based on the analysis of different variants of the names of suras and the need for ethnocultural adaptation of individual concepts in order to ensure the adequate perception of certain concepts by the inhabitants of the China. The problems of translating the Qurʼān into Chinese are largely related to the typological difference between the two linguistic systems. In this regard, the analysis of the translations of the names of suras of the Qurʼān into Chinese seems to be the most optimal starting point for the study of this topic.

Поэма А.А. Ахматовой «Реквием» в диалоге культур: проблемы перевода на осетинский язык

The article provides a comparative analysis of the poem by A.A. Akhmatova (1889–1966) “Requiem” in the original and in the translation of the Ossetian poet A.M. Kozachi (1937–2021) to identify similarities and differences in lexico-semantic means, figurative system and symbolism. The purpose of the study is to trace the representation of the semantic space of the original in poetic translation. The author shows the main types of translation strategies in the transfer of idiostyle features, gives an assessment from the position of reflecting the emotional message of the original in translation as a whole. The search for invariant correspondences in the target language showed the translator’s desire for more emotional identity with the original and partly for stylistic one. With the help of functional equivalents A.M. Kozachi managed to reflect in the translated text the personal experiences of the lyrical female character, to convey the confessional pathos of the poem. The absence of equivalents in the target language to convey some poetic images forced the translator to resort to associative substitution. The folklore, Orthodox symbolism found in the text is sometimes translated by non-literal lexical means. Since this or that symbol could not always evoke similar associations in the recipient, the translator resorted to semantic adaptation.

Point Cloud Registration Using Multiattention Mechanism and Deep Hybrid Features

Due to some unfavorable factors, how to accurately register point clouds is still a challenging task. In this article, an effective point cloud registration network is proposed with multiple attention mechanism and deep hybrid features. For the features obtained with a graph neural network, three attention modules, namely the spatial attention module, channel attention module, and self-geometric attention module, are utilized to mine various areas of regional information. An attention-based feature fusion module, which consists of three consecutive residual blocks, is devised to fuse the features from the three attention modules. Moreover, the capability of the network for correctly matching point clouds is enhanced, by using deep hybrid features to guide the correspondence search and the calculation of matching confidence. Experimental results on several widely used datasets demonstrate the effectiveness of the proposed point cloud registration network.

Construction of damage-free digital twin of damaged aero-engine blades for repair volume generation in remanufacturing

Accurate repair volume generation from 3D scan data of damaged aero-engine blades is of great importance in additive or hybrid remanufacturing for restoring the blades to a like-new condition. In addition to material-missing damages, the blade's surface also deforms due to working in harsh environments, which makes it deviate from the nominal design geometry. Therefore, the Boolean operation between the nominal CAD model and the scanned point cloud of the damaged blade does not yield an accurate repair volume. This paper presents a new methodology to construct an accurate damage-free digital twin model of the defective blades that contains the deformations of the blade's undamaged regions. The Boolean difference between the scan data of the damaged blade and its damage-free digital twin yields the repair volume with a smooth geometric transition at the interface of the repair patch and unrepaired regions. At first, the data points of damaged regions of the blade surface are detected and eliminated from the scan through a region growing segmentation. Then, a CAD-to-scan non-rigid registration algorithm deforms the nominal CAD model of the blade to best match it to the scanned point cloud in the undamaged regions. The non-rigid registration algorithm iteratively minimizes the distance between two datasets under the local rigidity constraint to avoid shrinkage and expansion of the deformed CAD model. A constrained point-to-surface weighted correspondence search method is proposed to reduce the influence of noise and unreliable correspondences on the non-rigid registration. The results of numerical and experimental case studies have demonstrated that the proposed method is accurate and robust to noise, and it can be effectively applied to construct a damage-free digital twin model for repair volume generation.

ShadowPatch: Shadow Based Segmentation for Reliable Depth Discontinuities in Photometric Stereo

AbstractPhotometric stereo is a well‐established method with outstanding traits to recover surface details and material properties, like surface albedo or even specularity. However, while the surface is locally well‐defined, computing absolute depth by integrating surface normals is notoriously difficult. Integration errors can be introduced and propagated by numerical inaccuracies from inter‐reflection of light or non‐Lambertian surfaces. But especially ignoring depth discontinuities for overlapping or disconnected objects, will introduce strong distortion artefacts. During the acquisition process the object is lit from different positions and self‐shadowing is in general considered as an unavoidable drawback, complicating the numerical estimation of normals. However, we observe that shadow boundaries correlate strongly with depth discontinuities and exploit the visual structure introduced by self‐shadowing to create a consistent image segmentation of continuous surfaces. In order to make depth estimation more robust, we deeply integrate photometric stereo with depth‐from‐stereo. Having obtained a shadow based segmentation of continuous surfaces, allows us to reduce the computational cost for correspondence search in depth‐from‐stereo. To speed‐up computation further, we merge segments into larger meta‐segments during an iterative depth optimization. The reconstruction error of our method is equal or smaller than previous work, and reconstruction results are characterized by robust handling of depth‐discontinuities, without any smearing artifacts.

Robust real-world point cloud registration by inlier detection

Real-world point cloud registration is challenging because of large outliers in correspondence search. The mixture variations, such as partial overlap, noise and cross sources, are the root cause of these large outliers. Existing methods face challenges in effectively removing the large outliers. We propose a novel coarse-to-fine framework to remove the outliers by detecting the accurate inlier correspondences. Specifically, our coarse module predicts the top-K accurate correspondences. The coarse module is trained by jointly leveraging global and local structured information. Then, our refinement module checks the correspondences further using our proposed novel higher-order filter, which enables the structure conformity of correspondences to improve the quality of inlier correspondences. The final transformation matrix is calculated by using the refined inlier correspondences. Furthermore, a new cross-source point cloud dataset is proposed to further demonstrate the robustness in real-world point clouds. Experimental results demonstrate that our algorithm achieves the state-of-the-art accuracy on both indoor and outdoor, same-source and newly proposed cross-source real-world point clouds.

Показания к первичной витрэктомии в лечении 3-й активной стадии ретинопатии недоношенных

Purpose is to determine the indications and evaluate the results of primary vitrectomy at 3rd stage of active ROP with an unfavorable type. Material and methods. A retrospective analysis of multimodal diagnostics was carried out in 16 premature infants from 2016 to 2019 who were treated with a diagnosis of 3rd active stage ROP, in which, due to the progression of the pathological process after transpupillary laser coagulation of the retina, vitreoretinal surgery was required during the period. In 2021, 3 patients (5 eyes) with an unfavorable type of 3rd active stages of ROP and confirmed extraretinal retinovitreal neovascularization underwent primary vitrectomy. Results. According to the results of a retrospective analysis, it was revealed that extraretinal retinovitreal neovascularization is a predictor of the development of retinal detachment. It is this clinical sign that indicates that transpupillary laser coagulation of the retina will be ineffective and will not lead to stabilization of the process, therefore, primary vitrectomy will be a pathogenetically justified treatment in such cases. Based on the data obtained in the type of a retrospective study in 2021, 3 patients with an unfavorable type of 3rd active stage ROP and confirmed extraretinal retinovitreal neovascularization underwent primary vitrectomy. After 3–6 months after the operation, the pathological process stabilized, the retina was completely attached, continued growth of vessels into the previously avascular zone of the retina was observed. Conclusion. To date, the search for a clear correspondence between a pathogenetically substantiated method of treatment and the timing of its implementation of the stage of the pathological process of ROP remains relevant. Extraretinal retinovitreal neovascularization according to multimodal preoperative diagnostics indicates the need for vitreretinal surgery as a pathogenetically substantiated method of treatment for 3rd active stage ROP. Keywords: retinopathy of prematurity, active stage, primary vitrectomy, transpupillary laser coagulation of the retina, early vitreous surgery.

Unsupervised Point Cloud Registration by Learning Unified Gaussian Mixture Models

Sampling noise and density variation widely exist in the point cloud acquisition process, leading to few accurate point-to-point correspondences. Since they rely on point-to-point correspondence search, existing state-of-the-art point cloud registration methods face difficulty in overcoming the sampling noise and density variation accurately or efficiently. Moreover, the recent state-of-the-art learning-based methods requires ground-truth transformation as supervised information which lead to large labor costs in real scenes. In this paper, our motivation is that two point-clouds are considered two samples from a unified Gaussian Mixture Model (UGMM). Then, we leverage the advantage of the statistic model to overcome the noise and density variants, and uses the alignment score in the UGMM to supervise the network training. To achieve this motivation, we propose a new unsupervised learning-based probabilistic registration algorithm to reconstruct the unified GMM and solve the registration problem simultaneously. The proposed method formulates the registration problem into a clustering problem, which estimates the posterior probability that classifies the points of two input point clouds to components of the unified GMM. A new feature interaction module is designed to learn the posterior probability using both the self and cross point cloud information. Then, two differential modules are proposed to calculate the GMM parameters and transformation matrices. Experimental results on synthetic and real-world point cloud datasets demonstrate that our unsupervised method achieves better registration accuracy and efficiency than the state-of-the-art supervised and semi-supervised methods in handling noisy and density variant point clouds.

Iterative Feature Matching for Self-Supervised Indoor Depth Estimation

In this paper, we propose an iterative feature matching framework for self-supervised depth estimation in indoor scenes. Conventional methods usually leverage the structure-from-motion supervision to help the photometric optimization escape from the local minima, which have complex ego-motion and large regions with non-texture or repeated-texture. However, the supervision is limited as the reconstruction is usually sparse. To address this, we propose an iterative feature matching framework called IFMNet to jointly learn depths and search for correspondences. With the predicted depths from the previous iteration, we present an online optimized grid searching algorithm to find more accurate correspondences. Given these new correspondences, we compute the triangulated depths and improve the depth network with adaptive bin-wise online hard example mining. Experimental results on the NYU Depth V2 and SceneNet datasets verify the effectiveness of our approach.

Хирургическое лечение ретинопатии недоношенных: исторический экскурс

Purpose. Тo study the change in approaches to surgical treatment of ROP in the historical aspect. Large-scale multicenter researches have begun in the world to study the mechanism of ROP development since 1951. For decades, knowledge about the pathogenesis of ROP has been expanded, confirmed in practice, and since the 80 s of the last century, the development of various methods for the treatment of active ROP, depending on the stage of the disease, has begun. It has been proven that cryo- and laser retinopexy of the avascular peripheral retina is an effective method of stabilizing the pathological process in the active stages. With the accumulation of experience, it became obvious that if after laser treatment it was not possible to achieve regression of the disease, and objective markers of its further progression are determined, then expectant tactics does not make sense. The search for a clear correspondence between a pathogenetically substantiated method of treatment and the timing of its implementation to the stage of the pathological process of ROP remains relevant nowadays. Conclusion. The history of the surgical treatment of ROP is the way of the development of knowledge about this disease. Currently, a promising direction in the management of patients with active PH is to shift the timing of treatment to earlier stages. Keywords: retinopathy of prematurity, surgical treatment, vitrectomy.

Exhaustive Search of Correspondences between Multimodal Remote Sensing Images Using Convolutional Neural Network.

Finding putative correspondences between a pair of images is an important prerequisite for image registration. In complex cases such as multimodal registration, a true match could be less plausible than a false match within a search zone. Under these conditions, it is important to detect all plausible matches. This could be achieved by an exhaustive search using a handcrafted similarity measure (SM, e.g., mutual information). It is promising to replace handcrafted SMs with deep learning ones that offer better performance. However, the latter are not designed for an exhaustive search of all matches but for finding the most plausible one. In this paper, we propose a deep-learning-based solution for exhaustive multiple match search between two images within a predefined search area. We design a computationally efficient convolutional neural network (CNN) that takes as input a template fragment from one image, a search fragment from another image and produces an SM map covering the entire search area in spatial dimensions. This SM map finds multiple plausible matches, locates each match with subpixel accuracy and provides a covariance matrix of localization errors for each match. The proposed CNN is trained with a specially designed loss function that enforces the translation and rotation invariance of the SM map and enables the detection of matches that have no associated ground truth data (e.g., multiple matches for repetitive textures). We validate the approach on multimodal remote sensing images and show that the proposed “area” SM performs better than “point” SM.