{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,14]],"date-time":"2026-08-14T15:56:52Z","timestamp":1786723012995,"version":"build-2736575974"},"reference-count":24,"publisher":"JMIR Publications Inc.","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["JMIR Med Inform"],"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:sec sec-type=\"background\">\n                    <jats:title>Background<\/jats:title>\n                    <jats:p>Deep learning models have shown strong potential for automated fracture detection in medical images. However, their robustness under varying image quality remains uncertain, particularly for small and subtle fractures, such as scaphoid fractures. Understanding how different types of image perturbations affect model performance is crucial for ensuring reliable deployment in clinical practice.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"objective\">\n                    <jats:title>Objective<\/jats:title>\n                    <jats:p>This study aimed to evaluate the robustness of a deep learning model trained to detect scaphoid fractures in radiographs when exposed to various image perturbations. We sought to identify which perturbations most strongly impact performance and to explore strategies to mitigate performance degradation.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"methods\">\n                    <jats:title>Methods<\/jats:title>\n                    <jats:p>Radiographic datasets were systematically modified by applying Gaussian noise, blurring, JPEG compression, contrast-limited adaptive histogram equalization, resizing, and geometric offsets. Model accuracy was evaluated across different perturbation types and levels. Image quality was quantified using peak signal-to-noise ratio and structural similarity index measure to assess correlations between degradation and model performance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"results\">\n                    <jats:title>Results<\/jats:title>\n                    <jats:p>Model accuracy declined with increasing perturbation severity, but the extent varied across perturbation types. Gaussian blur caused the most substantial performance drop, whereas contrast-limited adaptive histogram equalization increased the false-negative rate. The model demonstrated higher resilience to color perturbations than to grayscale degradations. A strong linear correlation was found between peak signal-to-noise ratio\u2013structural similarity index measure and accuracy, suggesting that better image quality led to improved detection. Geometric offsets and pixel value rescaling had minimal influence, whereas resolution was the dominant factor affecting performance.<\/jats:p>\n                  <\/jats:sec>\n                  <jats:sec sec-type=\"conclusions\">\n                    <jats:title>Conclusions<\/jats:title>\n                    <jats:p>The findings indicate that image quality, especially resolution and blurring, substantially influences the robustness of deep learning\u2013based fracture detection models. Ensuring adequate image resolution and quality control can enhance diagnostic reliability. These results provide valuable insights for designing more accurate and resilient medical imaging models under real-world variability.<\/jats:p>\n                  <\/jats:sec>","DOI":"10.2196\/65596","type":"journal-article","created":{"date-parts":[[2025,12,8]],"date-time":"2025-12-08T20:50:08Z","timestamp":1765227008000},"page":"e65596-e65596","source":"Crossref","is-referenced-by-count":1,"title":["Effects of Image Degradation on Deep Neural Network Classification of Scaphoid Fracture Radiographs: Comparison Study of Different Noise Types"],"prefix":"10.2196","volume":"14","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1079-7327","authenticated-orcid":false,"given":"Chihung","family":"Lin","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3908-4865","authenticated-orcid":false,"given":"Alfred P","family":"Yoon","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-5771-0340","authenticated-orcid":false,"given":"Chien-Wei","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0002-1707-0024","authenticated-orcid":false,"given":"Tung","family":"Chao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2822-3164","authenticated-orcid":false,"given":"Kevin C","family":"Chung","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-9770-5730","authenticated-orcid":false,"given":"Chang-Fu","family":"Kuo","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1010","published-online":{"date-parts":[[2026,1,22]]},"reference":[{"key":"R1","doi-asserted-by":"publisher","DOI":"10.2196\/55833","article-title":"Current state of community-driven radiological AI deployment in medical imaging","volume":"3","author":"Gupta","journal-title":"JMIR AI"},{"key":"R2","unstructured":"Hendrycks D Dietterich T . Benchmarking neural network robustness to common corruptions and perturbations. arXiv. Preprint posted online on Mar 28, 2019. [doi: 10.48550\/arXiv.1903.12261]"},{"key":"R3","doi-asserted-by":"crossref","unstructured":"Habibi Aghdam H Jahani Heravi E Puig D . Analyzing the stability of convolutional neural networks against image degradation. Presented at: International Conference on Computer Vision Theory and Applications; Feb 27-29, 2016:370-382; Rome, Italy. [doi: 10.5220\/0005720703700382]","DOI":"10.5220\/0005720703700382"},{"key":"R4","doi-asserted-by":"crossref","unstructured":"Hosseini H Xiao B Poovendran R . Google\u2019s cloud vision API is not robust to noise. Presented at: 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA); Dec 18-21, 2017; Cancun, Mexico. 2017.[doi: 10.1109\/ICMLA.2017.0-172]","DOI":"10.1109\/ICMLA.2017.0-172"},{"key":"R5","doi-asserted-by":"crossref","unstructured":"Dodge S Karam L . A study and comparison of human and deep learning recognition performance under visual distortions. Presented at: 2017 26th International Conference on Computer Communication and Networks (ICCCN); Jul 31 to Aug 3, 2017:1-7; Vancouver, BC, Canada. 2017.[doi: 10.1109\/ICCCN.2017.8038465]","DOI":"10.1109\/ICCCN.2017.8038465"},{"key":"R6","unstructured":"Fawzi A Moosavi-Dezfooli SM Frossard P . Robustness of classifiers: from adversarial to random noise. Presented at: NIPS\u201916: Proceedings of the 30th International Conference on Neural Information Processing Systems; Dec 5-10, 2016:1632-1640; Barcelona, Spain. [doi: 10.5555\/3157096.3157279]"},{"key":"R7","doi-asserted-by":"publisher","DOI":"10.2196\/66622","article-title":"Effectiveness of AI for enhancing computed tomography image quality and radiation protection in radiology: systematic review and meta-analysis","volume":"27","author":"Zhang","journal-title":"J Med Internet Res"},{"issue":"1","key":"R8","doi-asserted-by":"publisher","first-page":"12","DOI":"10.1007\/s10462-024-11005-9","article-title":"Robustness in deep learning models for medical diagnostics: security and adversarial challenges towards robust AI applications","volume":"58","author":"Javed","journal-title":"Artif Intell Rev"},{"issue":"1","key":"R9","doi-asserted-by":"publisher","DOI":"10.1038\/s41746-021-00393-9","article-title":"Recalibration of deep learning models for abnormality detection in smartphone-captured chest radiograph","volume":"4","author":"Kuo","journal-title":"NPJ Digit Med"},{"issue":"5","key":"R10","doi-asserted-by":"publisher","DOI":"10.1001\/jamanetworkopen.2021.6096","article-title":"Development and validation of a deep learning model using convolutional neural networks to identify scaphoid fractures in radiographs","volume":"4","author":"Yoon","journal-title":"JAMA Netw Open"},{"issue":"9","key":"R11","doi-asserted-by":"publisher","DOI":"10.3390\/s23094324","article-title":"Deep learning neural network performance on NDT digital x-ray radiography images: analyzing the impact of image quality parameters\u2014an experimental study","volume":"23","author":"Hena","journal-title":"Sensors (Basel)"},{"key":"#cr-split#-R12.1","doi-asserted-by":"crossref","unstructured":"Nazar\u00e9 TS da Costa GBP Contato W. A Ponti M . Deep convolutional neural networks and noisy images. Presented at: Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications. Nov 7-10, 2017","DOI":"10.1007\/978-3-319-75193-1_50"},{"key":"#cr-split#-R12.2","doi-asserted-by":"crossref","unstructured":"Springer International Publishing. Valpara\u00edso, Chile. 2018.[doi: 10.1007\/978-3-319-75193-1_50]","DOI":"10.1007\/978-3-319-75193-1_50"},{"key":"R13","unstructured":"Tan M Le Q . Efficientnet: rethinking model scaling for convolutional neural networks. Presented at: Proceedings of the 36th International Conference on Machine Learning; Jun 9-15, 2019:6105-6114; Long Beach, California, USA. 2019.[doi: 10.48550\/arXiv.1905.11946]"},{"issue":"4","key":"R14","doi-asserted-by":"publisher","first-page":"600","DOI":"10.1109\/tip.2003.819861","article-title":"Image quality assessment: from error visibility to structural similarity","volume":"13","author":"Wang","journal-title":"IEEE Trans Image Process"},{"key":"R15","unstructured":"Wang Z . The SSIM index for image quality assessment. Laboratory for Computational Vision, New York University. URL: http:\/\/www.cns.nyu.edu\/~lcv\/ssim [Accessed 23-12-2025]"},{"key":"R16","unstructured":"Vernon D . Machine Vision: Automated Visual Inspection and Robot Vision. Prentice-Hall; 1991. URL: https:\/\/archive.org\/details\/machinevisionaut0000vern [Accessed 23-12-2025]"},{"key":"R17","doi-asserted-by":"crossref","unstructured":"Dodge S Karam L . Understanding how image quality affects deep neural networks. Presented at: 2016 Eighth International Conference on Quality of Multimedia Experience (QoMEX); Jun 6-8, 2016:1-6; Lisbon, Portugal. [doi: 10.1109\/QoMEX.2016.7498955]","DOI":"10.1109\/QoMEX.2016.7498955"},{"key":"R18","unstructured":"Gonzalez RC Woods RE . Digital Image Processing. Addison-Wesley Publishing Company; 1992. ISBN: 9780201508031"},{"key":"R19","doi-asserted-by":"crossref","unstructured":"Zuiderveld K . Contrast limited adaptive histogram equalization. In: Heckbert PS , editor. Graphics Gems IV. Academic Press Professional Inc; 1994:474-485. [doi: 10.1016\/B978-0-12-336156-1.50061-6]","DOI":"10.1016\/B978-0-12-336156-1.50061-6"},{"issue":"3","key":"R20","doi-asserted-by":"publisher","first-page":"355","DOI":"10.1016\/S0734-189X(87)80186-X","article-title":"Adaptive histogram equalization and its variations","volume":"39","author":"Pizer","journal-title":"Comput Vis Graph Image Process"},{"key":"R21","doi-asserted-by":"crossref","unstructured":"Hor\u00e9 A Ziou D . Image quality metrics: PSNR vs. SSIM. Presented at: 2010 20th International Conference on Pattern Recognition; Aug 23-26, 2010:2366-2369; Istanbul, Turkey. [doi: 10.1109\/ICPR.2010.579]","DOI":"10.1109\/ICPR.2010.579"},{"key":"R22","doi-asserted-by":"publisher","DOI":"10.3389\/fradi.2023.1274273","article-title":"RoMIA: A framework for creating robust medical imaging AI models for chest radiographs","volume":"3","author":"Anand","journal-title":"Front Radiol"},{"key":"R23","doi-asserted-by":"publisher","DOI":"10.2196\/49949","article-title":"Democratizing artificial intelligence imaging analysis with automated machine learning: tutorial","volume":"25","author":"Thirunavukarasu","journal-title":"J Med Internet Res"}],"container-title":["JMIR Medical Informatics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/medinform.jmir.org\/2026\/1\/e65596","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,31]],"date-time":"2026-07-31T22:29:31Z","timestamp":1785536971000},"score":1,"resource":{"primary":{"URL":"https:\/\/medinform.jmir.org\/2026\/1\/e65596"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1,22]]},"references-count":24,"URL":"https:\/\/doi.org\/10.2196\/65596","relation":{},"ISSN":["2291-9694"],"issn-type":[{"value":"2291-9694","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,1,22]]},"article-number":"v14i1e65596"}}