{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,14]],"date-time":"2026-01-14T17:24:20Z","timestamp":1768411460151,"version":"3.49.0"},"reference-count":74,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2024,1,1]],"date-time":"2024-01-01T00:00:00Z","timestamp":1704067200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/legalcode"}],"funder":[{"name":"Bundesministerium f\u00fcr Bildung und Forschung (BMBF) Project Self-organizing Personal Knowledge Assistants in Evolving Corporate Memories","award":["BMBF 01IW20007"],"award-info":[{"award-number":["BMBF 01IW20007"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Access"],"published-print":{"date-parts":[[2024]]},"DOI":"10.1109\/access.2024.3425910","type":"journal-article","created":{"date-parts":[[2024,7,10]],"date-time":"2024-07-10T17:37:12Z","timestamp":1720633032000},"page":"103537-103564","source":"Crossref","is-referenced-by-count":5,"title":["The Reality of High Performing Deep Learning Models: A Case Study on Document Image Classification"],"prefix":"10.1109","volume":"12","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-3098-2458","authenticated-orcid":false,"family":"Saifullah","sequence":"first","affiliation":[{"name":"Smart Data and Knowledge Services (SDS), DFKI GmbH, Kaiserslautern, Germany"}]},{"given":"Stefan","family":"Agne","sequence":"additional","affiliation":[{"name":"Smart Data and Knowledge Services (SDS), DFKI GmbH, Kaiserslautern, Germany"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6100-8255","authenticated-orcid":false,"given":"Andreas","family":"Dengel","sequence":"additional","affiliation":[{"name":"Smart Data and Knowledge Services (SDS), DFKI GmbH, Kaiserslautern, Germany"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4239-6520","authenticated-orcid":false,"given":"Sheraz","family":"Ahmed","sequence":"additional","affiliation":[{"name":"Smart Data and Knowledge Services (SDS), DFKI GmbH, Kaiserslautern, Germany"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/ICDAR.2015.7333933"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/ICDAR.2017.149"},{"key":"ref3","volume-title":"DocXClassifier: Towards a Robust and Interpretable Deep Neural Network for Document Image Classification","author":"Saifullah","year":"2023"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-86331-9_47"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-50417-5_29"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.3390\/electronics8080832"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1613\/jair.1.13200"},{"key":"ref8","first-page":"1","article-title":"Imagenet-trained CNNs are biased towards texture; increasing shape bias improves accuracy and robustness","volume-title":"Proc. Int. Conf. Learn. Represent.","author":"Geirhos"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-12053-4_4"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1002\/widm.1356"},{"key":"ref11","article-title":"Characterising bias in compressed models","author":"Hooker","year":"2020","journal-title":"arXiv:2010.03058"},{"key":"ref12","article-title":"Shedding light on black box machine learning algorithms: Development of an axiomatic framework to assess the quality of methods that explain individual predictions","author":"Honegger","year":"2018","journal-title":"arXiv:1808.05054"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2018.2886017"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/sp.2017.49"},{"key":"ref15","first-page":"1","article-title":"Benchmarking neural network robustness to common corruptions and perturbations","volume-title":"Proc. Int. Conf. Learn. Represent.","author":"Hendrycks"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/ICDAR.2015.7333910"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.coling-main.82"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/ICPR56361.2022.9956167"},{"key":"ref19","volume-title":"Shabbypages: A Reproducible Document Denoising and Binarization Dataset","author":"Project","year":"2023"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1016\/j.culher.2017.10.001"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/N16-3020"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-019-01228-7"},{"key":"ref23","article-title":"SmoothGrad: Removing noise by adding noise","author":"Smilkov","year":"2017","journal-title":"arXiv:1706.03825"},{"key":"ref24","first-page":"4768","article-title":"A unified approach to interpreting model predictions","volume-title":"Proc. 31st Int. Conf. Neural Inf. Process. Syst.","author":"Lundberg"},{"key":"ref25","article-title":"Deep inside convolutional networks: Visualising image classification models and saliency maps","author":"Simonyan","year":"2013","journal-title":"arXiv:1312.6034"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10590-1_53"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1145\/3677119"},{"key":"ref28","first-page":"1488","article-title":"Countergan: Generating counterfactuals for real-time recourse and interpretability using residual GANs","volume-title":"Proc. 38th Conf. Uncertainty Artif. Intell.","volume":"180","author":"Nemirovsky"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1016\/j.patcog.2023.109666"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/icpr.2018.8545630"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1109\/ICDAR.2017.71"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3503161.3548112"},{"issue":"3","key":"ref33","first-page":"1","article-title":"Visualizing higher-layer features of a deep network","volume":"1341","author":"Erhan","year":"2009","journal-title":"Univeriste de Montreal"},{"key":"ref34","first-page":"3145","article-title":"Learning important features through propagating activation differences","volume-title":"Proc. 34th Int. Conf. Mach. Learn.","author":"Shrikumar"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.25300\/MISQ\/2014\/38.1.04"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1007\/s10044-021-01055-y"},{"key":"ref37","first-page":"673","article-title":"Explaining in style: Training a GAN to explain a classifier in styleSpace","volume-title":"Proc. IEEE\/CVF Int. Conf. Comput. Vis. (ICCV)","author":"Lang"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1109\/ICDAR.1995.601965"},{"key":"ref39","first-page":"606","article-title":"Document image retrieval based on layout structural similarity","volume-title":"Proc. Int. Conf. Image Process. Comput. Vis., Pattern Recognit.","volume":"2","author":"Shin"},{"key":"ref40","first-page":"653","article-title":"Structured document classification by matching local salient features","volume-title":"Proc. 21st Int. Conf. Pattern Recognit. (ICPR)","author":"Chen"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1016\/j.patrec.2013.10.030"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2003.1190578"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1007\/s10032-006-0020-2"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/ICPR.2014.546"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01167"},{"key":"ref46","first-page":"39","article-title":"Unidoc: Unified pretraining framework for document understanding","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"34","author":"Gu"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i04.6064"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0130140"},{"key":"ref49","first-page":"1097","article-title":"ImageNet classification with deep convolutional neural networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst. (NIPS)","author":"Krizhevsky"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/ACPR.2015.7486599"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.308"},{"key":"ref53","article-title":"Mobilenets: Efficient convolutional neural networks for mobile vision applications","author":"Howard","year":"2017","journal-title":"arXiv:1704.04861"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2017.243"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2019.2938758"},{"key":"ref56","first-page":"1059","article-title":"High-performance large-scale image recognition without normalization","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Brock"},{"key":"ref57","first-page":"6105","article-title":"EfficientNet: Rethinking model scaling for convolutional neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Tan"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.4324\/9781410605337-29"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref60","first-page":"1","article-title":"Striving for simplicity: The all convolutional net","volume-title":"Proc. ICLR","author":"Springenberg"},{"key":"ref61","first-page":"3319","article-title":"Axiomatic attribution for deep networks","volume-title":"Proc. Int. Conf. Mach. Learn.","volume":"70","author":"Sundararajan"},{"key":"ref62","article-title":"Not just a black box: Learning important features through propagating activation differences","author":"Shrikumar","year":"2017","journal-title":"arXiv:1605.01713"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2012.120"},{"key":"ref64","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2016.2599820"},{"key":"ref65","first-page":"1","article-title":"On the (in)fidelity and sensitivity of explanations","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"32","author":"Yeh"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.3390\/s22114084"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.5555\/1248547.1248548"},{"issue":"5","key":"ref68","first-page":"1","article-title":"Should we really use post-hoc tests based on mean-ranks?","volume":"17","author":"Benavoli","year":"2016","journal-title":"J. Mach. Learn. Res."},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/ICDARW.2019.10029"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/WACV48630.2021.00225"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-86549-8_10"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2021.3133200"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-57058-3_8"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1007\/s10032-023-00429-8"}],"container-title":["IEEE Access"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6287639\/10380310\/10591981.pdf?arnumber=10591981","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,8,7]],"date-time":"2024-08-07T06:35:18Z","timestamp":1723012518000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10591981\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024]]},"references-count":74,"URL":"https:\/\/doi.org\/10.1109\/access.2024.3425910","relation":{},"ISSN":["2169-3536"],"issn-type":[{"value":"2169-3536","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024]]}}}