{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,14]],"date-time":"2026-02-14T07:40:10Z","timestamp":1771054810300,"version":"3.50.1"},"reference-count":58,"publisher":"IOP Publishing","issue":"2","license":[{"start":{"date-parts":[[2025,4,22]],"date-time":"2025-04-22T00:00:00Z","timestamp":1745280000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"},{"start":{"date-parts":[[2025,4,22]],"date-time":"2025-04-22T00:00:00Z","timestamp":1745280000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"name":"National Science and Technology Council","award":["NSTC 112-2221-E-006-151-MY3"],"award-info":[{"award-number":["NSTC 112-2221-E-006-151-MY3"]}]},{"name":"Google Research Scholar Award"}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2025,6,30]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>For artificial intelligence applications in transmission electron microscopy (TEM), hardware and computational constraints often obstruct real-time data processing, inflating operational costs, consuming valuable instrument time, and heightening the risk of damage to beam-sensitive specimens, thereby complicating reliable data interpretation. To address these issues, we propose a two-stage pruning strategy that reduces deep-learning model size and computational overhead while preserving high performance and generalization across diverse datasets. Unlike conventional pruning techniques, which typically rely solely on weight magnitude and risk overlooking critical variability and directional properties in weight vectors, our approach initially removes filters with low magnitude and insufficient variability, followed by pruning filters with high linear similarity to eliminate redundancy. This one-shot pruning process, followed by fine-tuning, minimizes accuracy loss and mitigates barriers to deep learning integration in TEM workflows. Our method expedites TEM analysis, enabling more efficient, real-time, and cost-effective materials characterization. Additionally, this work lays a foundation for investigating the broader applicability and versatility to different architectures and tasks, particularly in resource-constrained environments where both model size and computational efficiency are critical.<\/jats:p>","DOI":"10.1088\/2632-2153\/adc9fa","type":"journal-article","created":{"date-parts":[[2025,4,7]],"date-time":"2025-04-07T22:54:53Z","timestamp":1744066493000},"page":"025021","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":2,"title":["Diversity-based two-phase pruning strategy for maximizing image segmentation generalization with applications in transmission electron microscopy"],"prefix":"10.1088","volume":"6","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-2046-0360","authenticated-orcid":true,"given":"Ze-Wei","family":"Ye","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-5944-4469","authenticated-orcid":false,"given":"Hung-Wei","family":"Hsueh","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0005-1258-3279","authenticated-orcid":true,"given":"Shu-Han","family":"Hsu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"266","published-online":{"date-parts":[[2025,4,22]]},"reference":[{"key":"mlstadc9fabib1","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/abd614","article-title":"Deep learning in electron microscopy","volume":"2","author":"Ede","year":"2021","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib2","doi-asserted-by":"publisher","first-page":"1896","DOI":"10.1017\/S1431927622012466","article-title":"Understanding the influence of receptive field and network complexity in neural network-guided TEM image analysis","volume":"28","author":"Sytwu","year":"2022","journal-title":"Microsc. Microanal."},{"key":"mlstadc9fabib3","doi-asserted-by":"publisher","first-page":"108","DOI":"10.1038\/s41524-020-00363-x","article-title":"Understanding important features of deep learning models for segmentation of high-resolution transmission electron microscopy images","volume":"6","author":"Horwath","year":"2020","journal-title":"npj Comput. Mater."},{"key":"mlstadc9fabib4","doi-asserted-by":"publisher","first-page":"239","DOI":"10.1111\/jmi.12716","article-title":"An overview of state\u2010of\u2010the\u2010art image restoration in electron microscopy","volume":"271","author":"Roels","year":"2018","journal-title":"J. Microsc."},{"key":"mlstadc9fabib5","doi-asserted-by":"publisher","first-page":"399","DOI":"10.1016\/j.micron.2004.02.003","article-title":"Radiation damage in the TEM and SEM","volume":"35","author":"Egerton","year":"2004","journal-title":"Micron"},{"key":"mlstadc9fabib6","first-page":"276","article-title":"De-noising filters for TEM (transmission electron microscopy) image of nanomaterials","author":"Kushwaha","year":"2012"},{"key":"mlstadc9fabib7","doi-asserted-by":"publisher","first-page":"675","DOI":"10.1126\/science.aao0865","article-title":"Atomic-resolution transmission electron microscopy of electron beam\u2013sensitive crystalline materials","volume":"359","author":"Zhang","year":"2018","journal-title":"Science"},{"key":"mlstadc9fabib8","doi-asserted-by":"publisher","first-page":"72","DOI":"10.1016\/j.micron.2019.01.005","article-title":"Radiation damage to organic and inorganic specimens in the TEM","volume":"119","author":"Egerton","year":"2019","journal-title":"Micron"},{"key":"mlstadc9fabib9","doi-asserted-by":"publisher","DOI":"10.1002\/smll.201906198","article-title":"A technical introduction to transmission electron microscopy for soft-matter: imaging, possibilities, choices, and technical developments","volume":"16","author":"Franken","year":"2020","journal-title":"Small"},{"key":"mlstadc9fabib10","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/acbb52","article-title":"Machine-learning approach for quantified resolvability enhancement of low-dose STEM data","volume":"4","author":"Gambini","year":"2023","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib11","doi-asserted-by":"publisher","first-page":"499","DOI":"10.1007\/BF03027093","article-title":"Artifacts in sample preparation of transmission electron microscopy","volume":"7","author":"Kim","year":"2001","journal-title":"Met. Mater. Int."},{"key":"mlstadc9fabib12","first-page":"125","article-title":"Artifacts in transmission electron microscopy","author":"Ayache","year":"2010"},{"key":"mlstadc9fabib13","doi-asserted-by":"publisher","first-page":"3708","DOI":"10.1007\/s11661-014-2331-0","article-title":"How TEM projection artifacts distort microstructure measurements: a case study in a 9 pct Cr-Mo-V steel","volume":"45","author":"Monsegue","year":"2014","journal-title":"Metall. Mater. Trans. A"},{"key":"mlstadc9fabib14","doi-asserted-by":"publisher","DOI":"10.1016\/j.commatsci.2024.113063","article-title":"Multi-phase material microscopic image segmentation for microstructure analysis of superalloys via modified U-Net and rectify strategies","volume":"242","author":"Zhou","year":"2024","journal-title":"Comput. Mater. Sci."},{"key":"mlstadc9fabib15","doi-asserted-by":"publisher","DOI":"10.1038\/srep16345","article-title":"Electron beam induced artifacts during in situ TEM deformation of nanostructured metals","volume":"5","author":"Sarkar","year":"2015","journal-title":"Sci. Rep."},{"key":"mlstadc9fabib16","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/acf6a9","article-title":"Combining variational autoencoders and physical bias for improved microscopy data analysis*","volume":"4","author":"Biswas","year":"2023","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib17","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ad1a4e","article-title":"Deep learning of crystalline defects from TEM images: a solution for the problem of \u2018never enough training data\u2019","volume":"5","author":"Govind","year":"2024","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib18","doi-asserted-by":"publisher","first-page":"36","DOI":"10.1038\/s41524-018-0093-8","article-title":"Automated defect analysis in electron microscopic images","volume":"4","author":"Li","year":"2018","journal-title":"npj Comput. Mater."},{"key":"mlstadc9fabib19","first-page":"1","article-title":"Instance segmentation of dislocations in TEM images","author":"Ruzaeva","year":"2023"},{"key":"mlstadc9fabib20","doi-asserted-by":"publisher","DOI":"10.1002\/adts.201800037","article-title":"A deep learning approach to identify local structures in atomic-resolution transmission electron microscopy images","volume":"1","author":"Madsen","year":"2018","journal-title":"Adv. Theor. Simul."},{"key":"mlstadc9fabib21","doi-asserted-by":"publisher","first-page":"17125","DOI":"10.1021\/acsnano.0c06809","article-title":"Statistical characterization of the morphologies of nanoparticles through machine learning based electron microscopy image analysis","volume":"14","author":"Lee","year":"2020","journal-title":"ACS Nano"},{"key":"mlstadc9fabib22","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ad073b","article-title":"Finding simplicity: unsupervised discovery of features, patterns, and order parameters via shift-invariant variational autoencoders*","volume":"4","author":"Ziatdinov","year":"2023","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib23","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ac3844","article-title":"Towards automating structural discovery in scanning transmission electron microscopy*","volume":"3","author":"Creange","year":"2022","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib24","doi-asserted-by":"publisher","first-page":"12488","DOI":"10.1021\/acsnano.5b05968","article-title":"High-throughput, algorithmic determination of nanoparticle structure from electron microscopy images","volume":"9","author":"Laramy","year":"2015","journal-title":"ACS Nano"},{"key":"mlstadc9fabib25","doi-asserted-by":"publisher","DOI":"10.1016\/j.xcrp.2022.100876","article-title":"Performance and limitations of deep learning semantic segmentation of multiple defects in transmission electron micrographs","volume":"3","author":"Jacobs","year":"2022","journal-title":"Cell Rep. Phys. Sci."},{"key":"mlstadc9fabib26","doi-asserted-by":"publisher","first-page":"454","DOI":"10.1016\/j.jechem.2021.12.001","article-title":"In situ transmission electron microscopy and artificial intelligence enabled data analytics for energy materials","volume":"68","author":"Zheng","year":"2022","journal-title":"J. Energy Chem."},{"key":"mlstadc9fabib27","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/ad9192","article-title":"From STEM-EDXS data to phase separation and quantification using physics-guided NMF","volume":"5","author":"Teurtrie","year":"2024","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstadc9fabib28","article-title":"Improving neural networks by preventing co-adaptation of feature detectors","author":"Hinton","year":"2012"},{"key":"mlstadc9fabib29","first-page":"818","article-title":"Visualizing and understanding convolutional networks","author":"Zeiler","year":"2014"},{"key":"mlstadc9fabib30","article-title":"Pruning filters for efficient ConvNets","author":"Li","year":"2017"},{"key":"mlstadc9fabib31","doi-asserted-by":"publisher","first-page":"118547","DOI":"10.1109\/ACCESS.2023.3326534","article-title":"V-SKP: vectorized kernel-based structured kernel pruning for accelerating deep convolutional neural networks","volume":"11","author":"Koo","year":"2023","journal-title":"IEEE Access"},{"key":"mlstadc9fabib32","doi-asserted-by":"publisher","first-page":"148","DOI":"10.1016\/j.neunet.2019.04.021","article-title":"Redundant feature pruning for accelerated inference in deep neural networks","volume":"118","author":"Ayinde","year":"2019","journal-title":"Neural Netw."},{"key":"mlstadc9fabib33","first-page":"824","article-title":"Leveraging filter correlations for deep model compression","author":"Singh","year":"2020"},{"key":"mlstadc9fabib34","first-page":"4335","article-title":"Filter pruning via geometric median for deep convolutional neural networks acceleration","author":"He","year":"2019"},{"key":"mlstadc9fabib35","first-page":"1475","article-title":"Compressing convolutional neural networks in the frequency domain","author":"Chen","year":"2016"},{"key":"mlstadc9fabib36","article-title":"Frequency-domain dynamic pruning for convolutional neural networks","volume":"vol 31","author":"Liu","year":"2018"},{"key":"mlstadc9fabib37","article-title":"Cnnpack: packing convolutional neural networks in the frequency domain","volume":"vol 29","author":"Wang","year":"2016"},{"key":"mlstadc9fabib38","first-page":"1","article-title":"Comparing different sequences of pruning algorithms for hybrid pruning","author":"Pragnesh","year":"2023"},{"key":"mlstadc9fabib39","doi-asserted-by":"publisher","first-page":"10558","DOI":"10.1109\/TPAMI.2024.3447085","article-title":"A survey on deep neural network pruning: taxonomy, comparison, analysis, and recommendations","volume":"46","author":"Cheng","year":"2024","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"mlstadc9fabib40","doi-asserted-by":"publisher","first-page":"370","DOI":"10.1016\/j.neucom.2021.07.045","article-title":"Pruning and quantization for deep neural network acceleration: a survey","volume":"461","author":"Liang","year":"2021","journal-title":"Neurocomputing"},{"key":"mlstadc9fabib41","first-page":"3398","article-title":"A simple hybrid filter pruning for efficient edge inference","author":"Basha","year":"2022"},{"key":"mlstadc9fabib42","doi-asserted-by":"publisher","first-page":"3161","DOI":"10.1007\/s10994-022-06193-w","article-title":"Pruning convolutional neural networks via filter similarity analysis","volume":"111","author":"Geng","year":"2022","journal-title":"Mach. Learn."},{"key":"mlstadc9fabib43","doi-asserted-by":"publisher","first-page":"533","DOI":"10.1016\/j.neucom.2021.08.098","article-title":"COP: customized correlation-based filter level pruning method for deep CNN compression","volume":"464","author":"Wang","year":"2021","journal-title":"Neurocomputing"},{"key":"mlstadc9fabib44","first-page":"11256","article-title":"Importance estimation for neural network pruning","author":"Molchanov","year":"2019"},{"key":"mlstadc9fabib45","article-title":"SqueezeNet: alexNet-level accuracy with 50x fewer parameters and <0.5 MB model size","author":"Iandola","year":"2016"},{"key":"mlstadc9fabib46","first-page":"1495","article-title":"Squeeze U-Net: a memory and energy efficient image segmentation network","author":"Beheshti","year":"2020"},{"key":"mlstadc9fabib47","doi-asserted-by":"publisher","first-page":"52804","DOI":"10.1109\/ACCESS.2022.3175188","article-title":"A squeeze U-SegNet architecture based on residual convolution for brain MRI segmentation","volume":"10","author":"Dayananda","year":"2022","journal-title":"IEEE Access"},{"key":"mlstadc9fabib48","first-page":"10096","article-title":"EfficientNetV2: smaller models and faster training","volume":"vol 139","author":"Tan","year":"2021"},{"key":"mlstadc9fabib49","first-page":"1473","article-title":"Eff-UNet: a novel architecture for semantic segmentation in unstructured environment","author":"Baheti","year":"2020"},{"key":"mlstadc9fabib50","first-page":"673","article-title":"Channel pruning via automatic structure search","author":"Lin","year":"2020"},{"key":"mlstadc9fabib51","first-page":"1","article-title":"Revisit kernel pruning with lottery regulated grouped convolutions","author":"Zhong","year":"2022"},{"key":"mlstadc9fabib52","first-page":"1526","article-title":"HRank: filter pruning using high-rank feature map","author":"Lin","year":"2020"},{"key":"mlstadc9fabib53","doi-asserted-by":"publisher","first-page":"175703","DOI":"10.1109\/ACCESS.2019.2957203","article-title":"Pruning blocks for CNN compression and acceleration via online ensemble distillation","volume":"7","author":"Wang","year":"2019","journal-title":"IEEE Access"},{"key":"mlstadc9fabib54","first-page":"2785","article-title":"Towards optimal structured CNN pruning via generative adversarial learning","author":"Lin","year":"2019"},{"key":"mlstadc9fabib55","first-page":"4510","article-title":"Resrep: lossless CNN pruning via decoupling remembering and forgetting","author":"Ding","year":"2021"},{"key":"mlstadc9fabib56","first-page":"2795","article-title":"Exploiting kernel sparsity and entropy for interpretable CNN compression","author":"Li","year":"2019"},{"key":"mlstadc9fabib57","first-page":"608","article-title":"Dhp: differentiable meta pruning via hypernetworks","author":"Li","year":"2020"},{"key":"mlstadc9fabib58","doi-asserted-by":"publisher","first-page":"90924","DOI":"10.1109\/ACCESS.2020.2993932","article-title":"Filter pruning without damaging networks capacity","volume":"8","author":"Zuo","year":"2020","journal-title":"IEEE Access"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,4,22]],"date-time":"2025-04-22T07:25:02Z","timestamp":1745306702000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/adc9fa"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,4,22]]},"references-count":58,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2025,4,22]]},"published-print":{"date-parts":[[2025,6,30]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/adc9fa","relation":{},"ISSN":["2632-2153"],"issn-type":[{"value":"2632-2153","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,4,22]]},"assertion":[{"value":"Diversity-based two-phase pruning strategy for maximizing image segmentation generalization with applications in transmission electron microscopy","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2025 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2025-01-09","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-04-07","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2025-04-22","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}