The traditional image classifiers are not capable to verify if samples belong to specified classes due to several reasons: classifiers do not provide boundaries between in-class and out-of-class samples; although classifiers provide separation boundaries between known classes, classifiers" latent features tend to have high intra-class variance; classifiers often predict high probabilities for out-of-distribution samples; training classifiers on unbalanced data results in bias towards over-represented classes. The nature of the class verification problem requires a different loss function than the ubiquitous cross entropy loss in traditional classifiers: input to a class verification function includes a suggested class in addition to an image. As opposed to outlier detection, space is transformed to be not only separable, but discriminative between in-class and out-of-class inputs. In this paper, class verification based on a euclidean distance from the class centre is proposed and implemented. Class centres are learnt by training on a centre loss function. The method"s effectiveness is shown on a self-checkout image dataset of 194 food retail products. The results show that a two-fold loss function is not only useful to verify class, but does not degrade classification performance - thus, the same neural network is usable both for classification and verification.
{"title":"Self-Checkout Product Class Verification using Center Loss approach","authors":"Bernardas Ciapas, P. Treigys","doi":"10.24132/csrn.3301.4","DOIUrl":"https://doi.org/10.24132/csrn.3301.4","url":null,"abstract":"The traditional image classifiers are not capable to verify if samples belong to specified classes due to several reasons: classifiers do not provide boundaries between in-class and out-of-class samples; although classifiers provide separation boundaries between known classes, classifiers\" latent features tend to have high intra-class variance; classifiers often predict high probabilities for out-of-distribution samples; training classifiers on unbalanced data results in bias towards over-represented classes. The nature of the class verification problem requires a different loss function than the ubiquitous cross entropy loss in traditional classifiers: input to a class verification function includes a suggested class in addition to an image. As opposed to outlier detection, space is transformed to be not only separable, but discriminative between in-class and out-of-class inputs. In this paper, class verification based on a euclidean distance from the class centre is proposed and implemented. Class centres are learnt by training on a centre loss function. The method\"s effectiveness is shown on a self-checkout image dataset of 194 food retail products. The results show that a two-fold loss function is not only useful to verify class, but does not degrade classification performance - thus, the same neural network is usable both for classification and verification.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"101 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121334818","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
W. Oronowicz-Jaśkowiak, Piotr Wasilewski, Mirosław Kowaluk
A technological interview was conducted with representatives of the judiciary in order to determine their expectations and beliefs related to the technological solution (involving detection of child sexual abuse materials using CNNs), being developed. The obtained results lead to the following conclusions: 1. Representatives of the judiciary recognize the advantages of the technological solution being created in the form of accelerating the work of experts and minimizing the risk of mistakes. 2. Representatives of the judiciary see the limitations of the technological solution being created in the form of the inability to replace court experts and emphasize that it also depends on the stage of the case. 3.The selection of pornographic materials from a specific set for later verification by a forensic expert is of the greatest importance.
{"title":"Justice Expectations Related to the Use of CNNs to Identify CSAM. Technological Interview","authors":"W. Oronowicz-Jaśkowiak, Piotr Wasilewski, Mirosław Kowaluk","doi":"10.24132/csrn.3301.39","DOIUrl":"https://doi.org/10.24132/csrn.3301.39","url":null,"abstract":"A technological interview was conducted with representatives of the judiciary in order to determine their expectations and beliefs related to the technological solution (involving detection of child sexual abuse materials using CNNs), being developed. The obtained results lead to the following conclusions: 1. Representatives of the judiciary recognize the advantages of the technological solution being created in the form of accelerating the work of experts and minimizing the risk of mistakes. 2. Representatives of the judiciary see the limitations of the technological solution being created in the form of the inability to replace court experts and emphasize that it also depends on the stage of the case. 3.The selection of pornographic materials from a specific set for later verification by a forensic expert is of the greatest importance.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"4 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117277209","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lung segmentation is an initial step to provide accurate lung parenchyma in many studies on lung diseases based on analyzing the Computed Tomography (CT) scan, especially in Non-Small Cell Lung Cancer (NSCLC) detection. In this work, Coordinate-UNet 3D, a model inspired by UNet, is proposed to improve the accuracy of lung segmentation in the CT scan. Like UNet, the proposed model consists of a contracting/encoder path to extract the high-level information and an expansive/decoder path to recover the features to provide the segmentation. However, we have considered modifying the structure inside each level of the model and using the Coordinate Convolutional layer as the final layer to provide the segmentation. This network was trained end-to-end from a small set of CT scans of NSCLC patients. The experimental results show the proposed network can provide a highly accurate segmentation for the validation set with a Dice Coefficient index of 0.991, an F1 score of 0.976, and a Jaccard index (IOU) of 0.9535.
{"title":"Coordinate-Unet 3D for segmentation of lung parenchyma","authors":"V. Le, Olivier Saut","doi":"10.24132/csrn.3301.6","DOIUrl":"https://doi.org/10.24132/csrn.3301.6","url":null,"abstract":"Lung segmentation is an initial step to provide accurate lung parenchyma in many studies on lung diseases based on analyzing the Computed Tomography (CT) scan, especially in Non-Small Cell Lung Cancer (NSCLC) detection. In this work, Coordinate-UNet 3D, a model inspired by UNet, is proposed to improve the accuracy of lung segmentation in the CT scan. Like UNet, the proposed model consists of a contracting/encoder path to extract the high-level information and an expansive/decoder path to recover the features to provide the segmentation. However, we have considered modifying the structure inside each level of the model and using the Coordinate Convolutional layer as the final layer to provide the segmentation. This network was trained end-to-end from a small set of CT scans of NSCLC patients. The experimental results show the proposed network can provide a highly accurate segmentation for the validation set with a Dice Coefficient index of 0.991, an F1 score of 0.976, and a Jaccard index (IOU) of 0.9535.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130108477","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
There are currently a range of different approaches for extracting iso-surfaces from volumetric medical image data. Of these, the HistoPyramid appears to be one of the more promising options. This is due to its use of stream compaction and expansion which facilitates extremely efficient traversal of the HistoPyramid structure. This paper introduces a novel extension to the HistoPyramid concept that entails incorporating a variable reduction between the HP layers in order to better fit volumes with arbitrary dimensions, thus saving memory and improving performance. As with the existing HistoPyramid techniques, the adaptive version lends itself to implementation on the GPU which in turn leads to further performance improvements. Ultimately, when compared against the best performing existing HistoPyramids, the adaptive approach yielded a performance improvement of up to 20 percent without any impact on the accuracy of the extracted mesh.
{"title":"Using the Adaptive HistoPyramid to Enhance Performance of Surface Extraction in 3D Medical Image Visualisation","authors":"Antony Padinjarathala, R. Sadleir","doi":"10.24132/csrn.3301.57","DOIUrl":"https://doi.org/10.24132/csrn.3301.57","url":null,"abstract":"There are currently a range of different approaches for extracting iso-surfaces from volumetric medical image data. Of these, the HistoPyramid appears to be one of the more promising options. This is due to its use of stream compaction and expansion which facilitates extremely efficient traversal of the HistoPyramid structure. This paper introduces a novel extension to the HistoPyramid concept that entails incorporating a variable reduction between the HP layers in order to better fit volumes with arbitrary dimensions, thus saving memory and improving performance. As with the existing HistoPyramid techniques, the adaptive version lends itself to implementation on the GPU which in turn leads to further performance improvements. Ultimately, when compared against the best performing existing HistoPyramids, the adaptive approach yielded a performance improvement of up to 20 percent without any impact on the accuracy of the extracted mesh.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126525787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Computational photography has revolutionized the way we capture and interpret images. Light fields, in particular, offer a rich representation of a scene"s geometry and appearance by encoding both spatial and angular information. In this paper, we present a novel approach to light field analysis that focuses on semantics. In contrast to the uniform distribution of samples in two-dimensional images, the distribution of samples in light fields varies for different scene regions. Some points are sampled from multiple directions, while others may only be captured by a small portion of the light field array. Our approach provides insights into this non-uniform distribution and helps guide further processing steps to fully leverage the available information content.
{"title":"SAIL: Semantic Analysis of Information in Light Fields: Results from Synthetic and Real-World Data","authors":"Robin Kremer, Thorsten Herfet","doi":"10.24132/csrn.3301.12","DOIUrl":"https://doi.org/10.24132/csrn.3301.12","url":null,"abstract":"Computational photography has revolutionized the way we capture and interpret images. Light fields, in particular, offer a rich representation of a scene\"s geometry and appearance by encoding both spatial and angular information. In this paper, we present a novel approach to light field analysis that focuses on semantics. In contrast to the uniform distribution of samples in two-dimensional images, the distribution of samples in light fields varies for different scene regions. Some points are sampled from multiple directions, while others may only be captured by a small portion of the light field array. Our approach provides insights into this non-uniform distribution and helps guide further processing steps to fully leverage the available information content.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133772756","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Edge bundling is one of the information visualization techniques, which bundle the edges of a network diagram based on certain rules to increase the visibility of the network diagram and facilitate the analysis of key relationships among nodes. As one of evolutionary-based edge bundling, genetic algorithm-based edge bundling (called GABEB) is proposed which uses a genetic algorithm to optimize the placement of edges based on aesthetic criteria. However, it does not sufficiently consider the bundling between neighboring edges, and thus visual clutter issues still remain. Based on the above background, we propose an improved bundling method that considers the concatenating of control points at each edge using GABEB.
{"title":"Evolutionary-Edge Bundling with Concatenation Process of Control Points","authors":"R. Saga, Jae-Chul Baek","doi":"10.24132/csrn.3301.33","DOIUrl":"https://doi.org/10.24132/csrn.3301.33","url":null,"abstract":"Edge bundling is one of the information visualization techniques, which bundle the edges of a network diagram based on certain rules to increase the visibility of the network diagram and facilitate the analysis of key relationships among nodes. As one of evolutionary-based edge bundling, genetic algorithm-based edge bundling (called GABEB) is proposed which uses a genetic algorithm to optimize the placement of edges based on aesthetic criteria. However, it does not sufficiently consider the bundling between neighboring edges, and thus visual clutter issues still remain. Based on the above background, we propose an improved bundling method that considers the concatenating of control points at each edge using GABEB.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"394 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115222994","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Roberto Soto-Villalobos, F. Benavides-Bravo, F. Hueyotl-Zahuantitla, Mario A. Aguirre-López
In this paper, we propose a new gasket fractal constructed in a deterministic iterated function system (IFS) way by means of interacting ball and square sets in R^2. The gasket consists of the ball sets generated by the IFS, possessing also exact self-similarity. All this leads to a direct deduction of other properties and a clear construction methodology, including a dynamic geometry procedure with an open-source construction protocol. We also develop an extended version of the fractal in R^n. Some resulting configurations consisting of stacked 2D-fractals are plotted. We discuss about potential applications of them in some areas of science, focusing mainly on percolation models. Guidelines for future work are also provided.
{"title":"A New Deterministic Gasket Fractal Based on Ball Sets","authors":"Roberto Soto-Villalobos, F. Benavides-Bravo, F. Hueyotl-Zahuantitla, Mario A. Aguirre-López","doi":"10.24132/csrn.3301.34","DOIUrl":"https://doi.org/10.24132/csrn.3301.34","url":null,"abstract":"In this paper, we propose a new gasket fractal constructed in a deterministic iterated function system (IFS) way by means of interacting ball and square sets in R^2. The gasket consists of the ball sets generated by the IFS, possessing also exact self-similarity. All this leads to a direct deduction of other properties and a clear construction methodology, including a dynamic geometry procedure with an open-source construction protocol. We also develop an extended version of the fractal in R^n. Some resulting configurations consisting of stacked 2D-fractals are plotted. We discuss about potential applications of them in some areas of science, focusing mainly on percolation models. Guidelines for future work are also provided.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123766547","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Miguel Seabra, Francisco Fernandes, Daniel Simões Lopes, João Madeiras Pereira
Interactive real-time rigid body simulation is a crucial tool in any modern game engine or 3D authoring tool. The quest for fast, robust and accurate simulations is ever evolving. PBRBD (Position Based Rigid Body Dynamics), a recent expansion of PBD (Position Based Dynamics), is a novel approach for this issue. This work aims at providing a comprehensible state-of-the art comparison between Position Based methods and other real-time simulation methods used for rigid body dynamics using a custom 3D physics engine for benchmarking and comparing PBRBD (Position Based Rigid Body Dynamics), and some variants, with state-of-the-art simulators commonly used in the gaming industry, PhysX and Havok. Showing that PBRBD can produce simulations that are accurate and stable, excelling at maintaining consistent energy levels, and allowing for a variety of constraints, but is limited in its handling of stable stacks of rigid bodies due to the propagation of rotational error.
{"title":"Position Based Rigid Body Simulation: A comparison of physics simulators for games","authors":"Miguel Seabra, Francisco Fernandes, Daniel Simões Lopes, João Madeiras Pereira","doi":"10.24132/csrn.3301.59","DOIUrl":"https://doi.org/10.24132/csrn.3301.59","url":null,"abstract":"Interactive real-time rigid body simulation is a crucial tool in any modern game engine or 3D authoring tool. The quest for fast, robust and accurate simulations is ever evolving. PBRBD (Position Based Rigid Body Dynamics), a recent expansion of PBD (Position Based Dynamics), is a novel approach for this issue. This work aims at providing a comprehensible state-of-the art comparison between Position Based methods and other real-time simulation methods used for rigid body dynamics using a custom 3D physics engine for benchmarking and comparing PBRBD (Position Based Rigid Body Dynamics), and some variants, with state-of-the-art simulators commonly used in the gaming industry, PhysX and Havok. Showing that PBRBD can produce simulations that are accurate and stable, excelling at maintaining consistent energy levels, and allowing for a variety of constraints, but is limited in its handling of stable stacks of rigid bodies due to the propagation of rotational error.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"14 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133676089","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
We introduce a Monte Carlo based real-time diffusion process for shape-based analysis in volumetric data. The diffusion process is carried out by using tiny massless particles termed shapetons, which are used to capture the shape information. Initially, these shapetons are randomly distributed inside the voxels of the volume data. The shapetons are then diffused in a Monte Carlo fashion to obtain the shape information. The direction of propagation for the shapetons is monitored by the Volume Gradient Operator (VGO). This operator is known for successfully capturing the shape information and thus the shape information is well captured by the shapeton diffusion method. All the shapetons are diffused simultaneously and all the results can be monitored in real-time. We demonstrate several important applications of our approach including colon cancer detection and design of shape-based transfer functions. We also present supporting results for the applications and show that this method works well for volumes. We show that our approach can robustly extract shape-based features and thus forms the basis for improved classification and exploration of features based on shape.
{"title":"Monte Carlo Based Real-Time Shape Analysis in Volumes","authors":"K. Gurijala, Lei Wang, A. Kaufman","doi":"10.24132/csrn.3301.15","DOIUrl":"https://doi.org/10.24132/csrn.3301.15","url":null,"abstract":"We introduce a Monte Carlo based real-time diffusion process for shape-based analysis in volumetric data. The diffusion process is carried out by using tiny massless particles termed shapetons, which are used to capture the shape information. Initially, these shapetons are randomly distributed inside the voxels of the volume data. The shapetons are then diffused in a Monte Carlo fashion to obtain the shape information. The direction of propagation for the shapetons is monitored by the Volume Gradient Operator (VGO). This operator is known for successfully capturing the shape information and thus the shape information is well captured by the shapeton diffusion method. All the shapetons are diffused simultaneously and all the results can be monitored in real-time. We demonstrate several important applications of our approach including colon cancer detection and design of shape-based transfer functions. We also present supporting results for the applications and show that this method works well for volumes. We show that our approach can robustly extract shape-based features and thus forms the basis for improved classification and exploration of features based on shape.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"252 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134374856","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The detection of mirrors is a challenging task due to their lack of a distinctive appearance and the visual similarity of reflections with their surroundings. While existing systems have achieved some success in mirror segmentation, the design of lightweight models remains unexplored, and datasets are mostly limited to clear mirrors in indoor scenes. In this paper, we propose a new dataset consisting of 454 images of outdoor mirrors and reflective surfaces. We also present a lightweight edge-guided convolutional neural network based on PMDNet. Our model uses EfficientNetV2-Medium as its backbone and employs parallel convolutional layers and a lightweight convolutional block attention module to capture both low-level and high-level features for edge extraction. It registered maximum F-measure scores of 0.8483, 0.8117, and 0.8388 on the Mirror Segmentation Dataset (MSD), Progressive Mirror Detection (PMD) dataset, and our proposed dataset, respectively. Applying filter pruning via geometric median resulted in maximum F-measure scores of 0.8498, 0.7902, and 0.8456, respectively, performing competitively with the state-of-the-art PMDNet but with 78.20x fewer floating-point operations per second and 238.16x fewer parameters. The code and dataset are available at https://github.com/memgonzales/mirror-segmentation.
{"title":"Designing a Lightweight Edge-Guided Convolutional Neural Network for Segmenting Mirrors and Reflective Surfaces","authors":"Mark Edward M. Gonzales, Lorene C. Uy, J. Ilao","doi":"10.24132/csrn.3301.14","DOIUrl":"https://doi.org/10.24132/csrn.3301.14","url":null,"abstract":"The detection of mirrors is a challenging task due to their lack of a distinctive appearance and the visual similarity of reflections with their surroundings. While existing systems have achieved some success in mirror segmentation, the design of lightweight models remains unexplored, and datasets are mostly limited to clear mirrors in indoor scenes. In this paper, we propose a new dataset consisting of 454 images of outdoor mirrors and reflective surfaces. We also present a lightweight edge-guided convolutional neural network based on PMDNet. Our model uses EfficientNetV2-Medium as its backbone and employs parallel convolutional layers and a lightweight convolutional block attention module to capture both low-level and high-level features for edge extraction. It registered maximum F-measure scores of 0.8483, 0.8117, and 0.8388 on the Mirror Segmentation Dataset (MSD), Progressive Mirror Detection (PMD) dataset, and our proposed dataset, respectively. Applying filter pruning via geometric median resulted in maximum F-measure scores of 0.8498, 0.7902, and 0.8456, respectively, performing competitively with the state-of-the-art PMDNet but with 78.20x fewer floating-point operations per second and 238.16x fewer parameters. The code and dataset are available at https://github.com/memgonzales/mirror-segmentation.","PeriodicalId":322214,"journal":{"name":"Computer Science Research Notes","volume":"80 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121126822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：