주요 콘텐츠로 건너뛰기

귀하의 브라우저가 완벽하게 지원되지 않습니다. 옵션이 있는 경우 최신 버전으로 업그레이드하거나 Mozilla Firefox, Microsoft Edge, Google Chrome 또는 Safari 14 이상을 사용하세요. 가능하지 않거나 지원이 필요한 경우 피드백을 보내주세요.

이 새로운 경험에 대한 귀하의 의견에 감사드립니다.의견을 말씀해 주세요(새 탭/창에서 열기)

Elsevier
엘스비어와 함께 출판
Connect

Five things every researcher should know about image manipulation

2020년 4월 29일 | 4분 읽기

저자: Kelechi Amakoh

image manipulation iStock-1158780306

© istockphoto.com/Tero Vesalainen

Important lessons from Researcher Academy’s latest webinar

Image manipulation poses a major challenge for the global research community and has prompted an ongoing discourse on ways to solve it. To advance this discourse and provide more context, Elsevier’s Researcher Academy(새 탭/창에서 열기) in partnership with the HEADT Centre at Humboldt University recently delivered a free(새 탭/창에서 열기) webinar on detecting image manipulation in scientific papers. This webinar takes a brief look at image manipulation, explores the limitations in curbing manipulation, showcases the tools currently being used in detecting such manipulations, and makes a case for an automated solution to the problem.

The webinar features a detailed presentation by a panel of three experts in the field of research integrity, image manipulation, and long-term digital archiving. The experts include Professor Dr. Michael Seadle, the founding co-director of the HEADT Centre; Dr. Thorsten Beck, a postdoctoral researcher at the centre and Elsevier’s Senior Vice President for Research Integrity, Dr. IJsbrand Jan Aalbersberg.

Below we present the five most important lessons from the image manipulation webinar.

  1. Image manipulation is still a serious issue in scholarship

    Image manipulation is a major challenge for the research community that requires a definitive solution. According to a short poll conducted by the presenters during the webinar, 52 percent of the audience rated image manipulation as a “very serious” challenge facing their field of research while an additional 36 percent rated it as “somewhat serious”. Only 18 percent saw image manipulation as not posing a serious problem in their field. Overall, the manipulation of scientific images is clearly a significant threat to reputable scholarship.

  2. Image "adjustment" is acceptable in some cases

    Minor adjustment/enhancement of images is acceptable in some cases. Permissible adjustment includes simple magnification of an image, and the addition of relevant arrows and highlights without distortion of the data. It must be said however, that when submitting research papers with images to journals for publication, researchers must ensure to adhere to publisher and journal-specific guidelines. You might find this page (새 탭/창에서 열기) useful to read more about what is/is not generally deemed acceptable.

  3. When acting as a peer reviewer, you should call out image manipulation if you see it

    It is important to challenge image manipulation as often as possible. Despite the lack of a one-stop solution to the problem, reviewers can at least provide evidence of aspects of an image that they think has been manipulated. By providing such concrete evidence, the editor can make an informed decision as to the future of the paper in question.

  4. Journals and editors have the final say about manipulated images

    Expert judgment over the issue of whether an image has been manipulated often vests in journal editors. In some cases, the institute at which the research took place might investigate and provide important evidence to the journal. Editors will be assisted in the assessment process by in-house rules, publisher-specific guidelines as well as the advice and input of the journal’s Editorial Board. It must be said that this is still very much a manual process and one which would certainly benefit from structural improvement.

  5. An automated solution is the ideal solution

    To reduce image manipulation in scholarship, it will be necessary to apply a degree of automation, although editors’ judgement will still be required to interpret the results. Automation would invest the process with a degree of scientific rigour and would save editors of journals and research integrity committees the headache of relying solely on visual inspection. Efforts are ongoing to achieve such a solution. However, many of the tools currently available for image manipulation detection are very much at the experimental stage and are not user-friendly. The various stakeholders will need to collaborate to ensure that such a solution can be realized within the shortest timeframe possible.

To learn more about image manipulation and how to avoid errors when preparing your images for publication, listen to Seadle and colleagues in this module: Detecting Image Manipulation: Routines, Tools & Limitations(새 탭/창에서 열기).

You can also check the frequently-asked-questions on image manipulation section for further answers by experts.

기여자

Image of Kelechi Amakoh

KA

Kelechi Amakoh

Marketing, communications and digital content

Researcher Academy