Elsevier Graphical Abstract Guidelines: Size, Format, and AI Policy

Last reviewed: July 12, 2026 · Reviewed by the SciFig Editorial Team

Elsevier publishes general guidance for graphical abstracts, but the exact journal's Guide for Authors remains the final authority. Some Elsevier journals require a graphical abstract, some make it optional, and some use different terminology or specifications.

This article summarizes Elsevier's current graphical abstract guidance and generative AI policies for journals. Both technical and AI requirements can change, so verify the official pages and target journal immediately before submission.

Elsevier general specifications

The current general guidance states:

• Minimum image size: 1328 × 531 pixels, width × height
• Minimum resolution: 300 dpi
• Larger images: retain the same 500:200 ratio
• Preferred file types: TIFF, EPS, PDF, or Microsoft Office files
• Reading order: a clear start and end, preferably top to bottom or left to right
• Fonts: Times, Arial, Courier, or Symbol, large enough to remain legible after reduction
• Submission: upload the graphical abstract as a separate file in the online submission system

Elsevier notes that the image is scaled proportionally to fit a 500 × 200 pixel window on ScienceDirect. Design for that reduced display, not only for the large working canvas.

1. Check the exact journal before designing

Do not start from the 1328 × 531 specification alone. Open the journal's Guide for Authors and record:

• Whether a graphical abstract is required, optional, or not accepted.
• Whether the article type is eligible.
• Any journal-specific canvas, caption, font, or file requirements.
• How the file should be named and categorized in the submission system.
• Current image, copyright, permissions, and AI policies.

If the journal instructions conflict with the general Elsevier page, follow the journal instructions or ask the editorial office.

2. Reduce the paper to one supported message

The purpose is to help readers grasp the take-home message and decide whether the article is relevant. It is not to reproduce the paper's figures or encode every result.

Write a one-sentence message containing:

1. The research subject or system.
2. The intervention, method, or central relationship.
3. The main outcome supported by the data.

Remove secondary results, long definitions, detailed methods, and speculative mechanisms unless they are essential. Keep observed evidence visually distinct from proposed interpretation.

3. Create an obvious reading order

Elsevier recommends a clear start and end, preferably left to right or top to bottom. Use spatial hierarchy before relying on arrows:

• Place the starting condition at the left or top.
• Keep the key mechanism or method near the visual center.
• Place the supported outcome at the right or bottom.
• Limit crossing connectors and decorative detours.
• Use consistent shapes and colors for the same type of entity.

The reader should not need to guess which arrow to follow first.

4. Design for the 500 × 200 display window

Even a 1328 × 531 source image will be reduced substantially online. Preview the artwork at 500 × 200 pixels and check:

• Labels remain readable.
• Arrowheads and inhibitory lines remain distinct.
• Thin structures do not disappear.
• Similar colors remain distinguishable.
• The core message remains visible without zooming.

Avoid unnecessary whitespace because it reduces the usable area after proportional scaling. At the same time, do not fill every gap; negative space is necessary to separate information groups.

5. Verify accuracy, rights, and originality

The authors remain responsible for the scientific and legal integrity of the image.

• Compare every label, structure, arrow, condition, and outcome with the manuscript and source data.
• Do not use microscopy, histology, western blots, radiology images, or other primary research images as decorative material.
• Confirm permission for third-party icons, photos, figures, and adapted artwork.
• Retain editable source files, licenses, attribution records, and review notes.
• Ask a coauthor familiar with the methods to conduct a final independent check.

6. Understand Elsevier's current AI rule

Elsevier's generative AI policy currently distinguishes several image categories. It allows limited, disclosed use for certain explanatory images, but specifically states that general-purpose generative AI image tools must not be used to create graphical abstracts. The policy recommends dedicated scientific illustration or professional illustration tools with appropriate licensing.

That means SciFig can support internal ideation and composition exploration, but its generated image should not be submitted as an Elsevier graphical abstract under the current policy. Rebuilding labels or lightly editing the output does not automatically make the workflow compliant.

AI output can also invent structures, labels, relationships, and references. Never treat it as research evidence. Do not upload confidential manuscripts, unpublished images, identifiable patient information, or proprietary data unless the service's terms and your institutional approvals allow that processing.

If your workflow includes any AI-assisted step, check the current policy for disclosure requirements and ask the journal when the category is uncertain.

7. Export and preflight

Before uploading the separate graphical-abstract file:

• [ ] Target journal and article-type instructions checked
• [ ] Minimum 1328 × 531 pixels and 300 dpi, unless the journal specifies otherwise
• [ ] Correct 500:200 ratio retained for a larger export
• [ ] Preferred accepted format used
• [ ] Reading path is obvious from left to right or top to bottom
• [ ] Fonts and key marks remain legible at 500 × 200 pixels
• [ ] All scientific content reviewed against the manuscript and data
• [ ] Third-party permissions, licenses, and attributions documented
• [ ] Current Elsevier and journal-specific AI policy checked
• [ ] File uploaded separately and categorized correctly

Frequently asked questions

Is 1328 × 531 pixels required for every Elsevier journal?

It is Elsevier's current general minimum, not a guarantee for every journal and article type. Always open the exact journal's Guide for Authors.

Can I submit a 4K graphical abstract?

A larger file can be acceptable if it preserves the required ratio, but more pixels do not fix poor hierarchy, illegible labels, incorrect science, unsupported claims, or a prohibited production method.

Can I use SciFig to make an Elsevier graphical abstract?

Under Elsevier's current policy, general-purpose generative AI image tools must not be used to create graphical abstracts. Use SciFig only for internal ideation, then follow the journal's permitted illustration workflow and verify all content.

Does 300 dpi guarantee good quality?

No. DPI must be considered together with actual pixel dimensions, physical size, compression, line quality, and final-size legibility. Changing metadata alone does not add detail.

Related SciFig resources

• Graphical abstract size and dimensions
• How to make a graphical abstract
• Graphical abstract examples
• Graphical Abstract Generator for internal concept exploration only

Official sources

• Elsevier: Graphical abstract in Elsevier journals
• Elsevier: Generative AI policies for journals
• Elsevier: Artwork and media instructions

Policy note: This independent guide is not affiliated with or endorsed by Elsevier. Official Elsevier policy and the target journal's current Guide for Authors always take precedence.