07342nam 22007935 450 99654685190331620230819122426.03-031-40773-310.1007/978-3-031-40773-4(MiAaPQ)EBC30713759(Au-PeEL)EBL30713759(DE-He213)978-3-031-40773-4(PPN)272260169(EXLCZ)992800503570004120230819d2023 u| 0engurcnu||||||||txtrdacontentcrdamediacrrdacarrierReproducible Research in Pattern Recognition[electronic resource] Fourth International Workshop, RRPR 2022, Montreal, Canada, August 21, 2022, Revised Selected Papers /edited by Bertrand Kerautret, Miguel Colom, Adrien Krähenbühl, Daniel Lopresti, Pascal Monasse, Benjamin Perret1st ed. 2023.Cham :Springer Nature Switzerland :Imprint: Springer,2023.1 online resource (127 pages)Lecture Notes in Computer Science,1611-3349 ;14068Print version: Kerautret, Bertrand Reproducible Research in Pattern Recognition Cham : Springer,c2023 9783031407727 Intro -- Preface -- Organization -- The Fuzzy Boundaries of Reproducibility (Lightweight Presentation Abstract) -- Contents -- Reproducible Research Framework -- Development Efforts for Reproducible Research: Platform, Library and Editorial Investment -- 1 Introduction -- 2 Reproducible Research Platform Updates -- 3 Reproducible Research Through Libraries -- 3.1 Library Experiences from Pattern Recognition, Image and Geometry Domains -- 3.2 Higra Library Development Feedback -- 4 Advanced Editorial Efforts -- 4.1 Improvements in the IPOL Journal -- 4.2 OVD-SaaS, a Spin-Off of IPOL for Industrial Applications -- 5 Conclusion -- References -- Reproducible Research Results -- Enhancing GNN Feature Modeling for Document Information Extraction Using Transformers -- 1 Introduction -- 2 Related Works -- 3 Proposed Model -- 3.1 Texts and Bounding Boxes -- 3.2 Features Assignment -- 3.3 Graph Construction -- 3.4 GNN Model -- 3.5 Model Prediction -- 4 Experiments -- 4.1 Dataset -- 4.2 Experimental Setup -- 4.3 Metrics -- 4.4 Results -- 4.5 Implementation Details -- 5 Conclusion -- References -- Short ICPR Companion Papers -- A Novel Pattern-Based Edit Distance for Automatic Log Parsing: Implementation and Reproducibility Notes -- 1 Introduction -- 2 Implementation Considerations -- 3 Installation Steps -- 4 Pattern Clustering Usage -- 4.1 Pattern Collection -- 4.2 Returned Value -- 4.3 Dropping Duplicated Pattern Automata -- 5 Experimental Setup -- 5.1 Drain and LogMine Integration -- 5.2 Loghub Dataset -- 5.3 Ground Truth -- 5.4 Experimental Parameters -- 5.5 Accuracy -- 6 Conclusion -- References -- Companion Paper: Deep Saliency Map Generators for Multispectral Video Classification -- 1 Introduction -- 2 Deep Saliency Map Generators -- 2.1 Grad-CAM -- 2.2 RISE -- 2.3 SIDU -- 3 Networks -- 3.1 3D-ResNet -- 3.2 Persistent Appearance Network.4 Evaluation -- 4.1 Deletion Metric -- 4.2 Insertion Metric -- 5 Conclusion -- References -- On Challenging Aspects of Reproducibility in Deep Anomaly Detection -- 1 Introduction -- 2 Deep Anomaly Detection -- 3 Challenges for Reproducibility -- 3.1 Nondeterminism in Network Optimization -- 3.2 Sensitivity to Hyperparameters -- 3.3 Complexity -- 3.4 Dataset Selection -- 3.5 Resource Limitations -- 3.6 Dependencies -- 4 Complexity-Evidence Tradeoff -- 5 Conclusion -- References -- On the Implementation of Baselines and Lightweight Conditional Model Extrapolation (LIMES) Under Class-Prior Shift -- 1 Introduction -- 2 Dataset -- 3 Implementation -- 3.1 Preprocessing of Raw Json Files with Twitter Data -- 3.2 Embeddings -- 3.3 Machine Learning Models - Training and Evaluation -- 3.4 Running Experiments Efficiently -- 4 Reproducibility -- 5 Credibility of Results -- 6 Conclusions -- References -- Special Reproducibility Track from DGMM Event -- Combining Max-Tree and CNN for Segmentation of Cellular FIB-SEM Images -- 1 Introduction -- 2 State of the Art -- 3 Methods -- 3.1 Max-Tree -- 3.2 Segmentation -- 3.3 Evaluation Metrics -- 4 Experiments -- 4.1 Data -- 4.2 Results -- 4.3 Reproducibility -- 5 Conclusion -- A Appendix -- A.1 Results -- A.2 Example Preprocessing Visualization -- References -- Automatic Forest Road Extraction from LiDAR Data Using Convolutional Neural Networks*-12pt -- 1 Introduction -- 2 Method -- 2.1 Problem Statement -- 2.2 Previous Approach to Forest Road Extraction -- 2.3 Light DDCM-Net Architecture -- 3 Experimental Setup -- 3.1 Dataset -- 3.2 Network Training -- 4 Results and Discussions -- 5 Conclusion -- References -- Discussions Report Paper -- Promoting Reproducibility of Research Results in International Events (Report from the 4th RRPR)*-12pt -- 1 Introduction -- 2 Addressing RR at International Conferences.2.1 Recent Proposals -- 2.2 New Ideas on Promoting RR at International Conferences -- 2.3 Impact of Efforts Encouraging RR in Conferences -- 3 Focus on Motivating RR -- 3.1 Recent Initiatives -- 3.2 Issues for Research Result Comparisons -- 3.3 Strengthening Reproducibility: From Publications to Teaching -- 4 Conclusion -- References -- Author Index.This book constitutes the thoroughly refereed post-workshop proceedings of the 4th International Workshop on Reproducible Research in Pattern Recognition, RRPR 2022, held in Montreal, Canada, in August 2022. The 5 revised full papers presented together with 4 short papers, were carefully reviewed and selected from 9 submissions. The papers were organized into three main categories.Lecture Notes in Computer Science,1611-3349 ;14068Application softwareComputer engineeringComputer networksComputersArtificial intelligenceComputers, Special purposeComputer and Information Systems ApplicationsComputer Engineering and NetworksComputing MilieuxArtificial IntelligenceComputer Communication NetworksSpecial Purpose and Application-Based SystemsApplication software.Computer engineering.Computer networks.Computers.Artificial intelligence.Computers, Special purpose.Computer and Information Systems Applications.Computer Engineering and Networks.Computing Milieux.Artificial Intelligence.Computer Communication Networks.Special Purpose and Application-Based Systems.005.3Kerautret Bertrand1423672Colom Miguel1423673Krähenbühl Adrien1423674Lopresti Daniel899206Monasse Pascal506784Perret Benjamin1423675MiAaPQMiAaPQMiAaPQBOOK996546851903316Reproducible Research in Pattern Recognition3552051UNISA