From Image Access to Research Reuse: IIIF for the Natural and Engineering Sciences

At the 2026 Inter­na­tion­al Image Inter­op­er­abil­i­ty Frame­work (IIIF) Annu­al Con­fer­ence, 4TU.ResearchData pre­sent­ed its work on using IIIF for researchers in the Nat­ur­al and Engi­neer­ing Sci­ences. The main mes­sage was sim­ple: IIIF should not be seen only as a visu­al­i­sa­tion lay­er for dig­i­tal col­lec­tions. For research repos­i­to­ries such as 4TU.ResearchData, IIIF can become part of the infra­struc­ture that enables sci­en­tif­ic reuse, inter­op­er­abil­i­ty, and com­pu­ta­tion­al access to image-based research data.

4TU.ResearchData is a data and soft­ware repos­i­to­ry for sci­ence, engi­neer­ing, and design. Its mis­sion is not only to pre­serve research out­puts, but also to sup­port their reuse in prac­ti­cal, scal­able, and inter­op­er­a­ble ways.

Use case 1: Turning scientific images into reusable web resources

Image-based datasets in the Nat­ur­al and Engi­neer­ing Sci­ences can be large, mul­ti-res­o­lu­tion, and com­pu­ta­tion­al­ly demand­ing. A researcher may find a valu­able microscopy dataset, satel­lite image col­lec­tion, engi­neer­ing scan, or his­tor­i­cal image train­ing set, but the reuse work­flow often still requires down­load­ing sev­er­al giga­bytes of data before any inspec­tion or analy­sis can begin. This work­flow is pos­si­ble, but it does not scale well for inter­ac­tive reuse.

By imple­ment­ing the IIIF Image API, 4TU.ResearchData can serve image regions, thumb­nails, and scaled ver­sions of large images on demand. Researchers do not need to down­load the full orig­i­nal file just to inspect, com­pare, pre­view, or analyse a spe­cif­ic part of an image. Instead, they can access the image through stan­dard­ised web requests.

This changes the role of the repos­i­to­ry. It is no longer only a place where files are deposit­ed and down­loaded. It becomes an inter­ac­tive image deliv­ery ser­vice that allows researchers, view­ers, note­books, scripts, and exter­nal plat­forms to request only the visu­al data they need.

Use case 2: Enriching IIIF manifests with repository metadata

Anoth­er use case pre­sent­ed at the con­fer­ence was the enrich­ment of IIIF man­i­fests with meta­da­ta already avail­able in the repos­i­to­ry. IIIF man­i­fests expose images in a stan­dard­ised way, but the meta­da­ta includ­ed in them is often fixed or lim­it­ed. This can make it dif­fi­cult for researchers to under­stand the geo­graph­ic con­text, tem­po­ral con­text, rela­tion­ship to pub­li­ca­tions, exper­i­men­tal back­ground, or domain-spe­cif­ic mean­ing of the images.

4TU.ResearchData already stores rich repos­i­to­ry meta­da­ta. Through the repos­i­to­ry API, this meta­da­ta can be retrieved and added to the IIIF man­i­fest. This allows the man­i­fest to expose more of the research con­text around the images.

A con­crete exam­ple is geospa­tial meta­da­ta. When a dataset con­tains loca­tion infor­ma­tion, this can be added to the IIIF man­i­fest in a way that exter­nal tools can under­stand. Tools such as Allmaps can then use this infor­ma­tion for geo­ref­er­enc­ing and map-based visu­al­i­sa­tion. In this sce­nario, the man­i­fest becomes a bridge between repos­i­to­ry meta­da­ta, exter­nal tools, and new research or out­reach work­flows.

This is inter­op­er­abil­i­ty in prac­tice. The repos­i­to­ry does not need to build every pos­si­ble tool itself. By expos­ing data and meta­da­ta through open stan­dards, 4TU.ResearchData can make deposits reusable in tools that already exist.

Why invest in IIIF?

Invest­ing in IIIF is not only a tech­ni­cal improve­ment. It sup­ports and aligns with sev­er­al strate­gic pri­or­i­ties for 4TU.ResearchData.

First, it strength­ens open sci­ence. FAIR data and soft­ware should not only be find­able and down­load­able, but also oper­a­tional­ly reusable. For large sci­en­tif­ic images, reuse requires more than a down­load but­ton. It requires stan­dard­ised access, machine-action­able descrip­tions, and inte­gra­tion with research work­flows.

Sec­ond, it sup­ports dig­i­tal sov­er­eign­ty. By imple­ment­ing IIIF with­in the repos­i­to­ry infra­struc­ture, 4TU.ResearchData keeps con­trol over image stor­age, APIs, meta­da­ta, access poli­cies, ren­der­ing, caching, authen­ti­ca­tion, and preser­va­tion work­flows. This reduces depen­dence on pro­pri­etary visu­al­i­sa­tion plat­forms or closed image-host­ing ecosys­tems.

Third, it enables inter­op­er­abil­i­ty with­out lock-in. IIIF is an inter­na­tion­al open stan­dard used by libraries, muse­ums, archives, uni­ver­si­ties, and research infra­struc­tures. By adopt­ing it, 4TU.ResearchData makes sci­en­tif­ic images com­pat­i­ble with a wider ecosys­tem of view­ers, anno­ta­tion tools, man­i­fest edi­tors, geo­ref­er­enc­ing plat­forms, and com­pu­ta­tion­al work­flows.

Final­ly, it sup­ports fed­er­at­ed research infra­struc­ture. Researchers should be able to com­pare, aggre­gate, anno­tate, and analyse image-based data across insti­tu­tions while the data remains gov­erned and pre­served by the repos­i­to­ry that hosts it. IIIF con­tributes to this vision by allow­ing dis­trib­uted image col­lec­tions to be accessed through shared pro­to­cols.

Challenges and next steps

There are still tech­ni­cal and meta­da­ta chal­lenges to solve. For exam­ple, zipped deposits can bypass IIIF pro­cess­ing, which means that images inside com­pressed archives may not become direct­ly acces­si­ble through IIIF ser­vices. File-lev­el meta­da­ta is also essen­tial. Dataset-lev­el meta­da­ta is use­ful, but many sci­en­tif­ic work­flows require meta­da­ta per image, per frame, per loca­tion, or per exper­i­men­tal con­di­tion.

These chal­lenges show why con­tin­ued invest­ment is need­ed. IIIF pro­vides a foun­da­tion for rich­er repos­i­to­ry ser­vices, includ­ing bet­ter pre­views, enriched man­i­fests, com­pu­ta­tion­al reuse, anno­ta­tion work­flows, geo­ref­er­enc­ing, out­reach col­lec­tions, and cross-repos­i­to­ry com­par­i­son.

The IIIF com­mu­ni­ty is an impor­tant part of this process. By par­tic­i­pat­ing in this com­mu­ni­ty, 4TU.ResearchData can learn from mature imple­men­ta­tions, reuse exist­ing tools, and con­tribute the per­spec­tive of sci­en­tif­ic data repos­i­to­ries.

For 4TU.ResearchData, IIIF is not only about improv­ing access to images. It is about mov­ing from image access to research reuse.

Cov­er image: AI gen­er­at­ed