Digital pathology has emerged as a critical domain for computer vision research, driven by the increasing availability of Whole Slide Images (WSIs) and the need for scalable, accurate, and interpretable diagnostic tools. At the same time, foundation models — including large-scale Vision Transformers (ViTs) and multimodal architectures — have redefined the state of the art in visual understanding tasks.
Despite these advances, several open challenges remain when applying foundation models to digital pathology:
Scalability to gigapixel images: WSIs require efficient tiling, aggregation, and hierarchical modeling strategies.
Limited labeled data: Annotation requires expert pathologists, motivating self-supervised, weakly supervised, and transfer learning approaches.
Domain adaptation and generalization: Variability in staining protocols and acquisition conditions impacts model robustness.
Multimodal learning: Integration of imaging data with clinical, genomic, and textual information remains underexplored.
Explainability and validation: Clinical adoption requires interpretable outputs and rigorous validation protocols.
Deployment constraints: Integration into clinical workflows demands efficient, reliable, and certifiable systems.
This workshop aims to provide a focused forum within SIBGRAPI for discussing methodological advances, benchmark datasets, and real-world applications, fostering interaction between academia, healthcare institutions, and industry.
More information: pathfm2026chairs@gmail.com (Organizing Committee).
The workshop will follow standard SIBGRAPI procedures and formatting guidelines to ensure consistency with the main conference.
All manuscripts submitted to this workshop must be in English and undergo double-blind peer review. The LaTeX template is available here.
Please follow the included instructions and the formatting guidelines, avoiding any change to the format (in particular, avoiding space command or creating subsections or paragraph titles with any command different from subsection or paragraph). The references should be formatted accordingly. Please do not change the font size, even if it helps meet the page limit. The illustrations in the paper should be generated either as true vectors or rasterized at least 300 dpi.
Paper submission will be handled via the CMT system available here.
The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.
The review process of all submissions will be double-blind. The authors’ identities will be tracked only by the submission system and visible only by the track chairs. The program committee members and reviewers will not know the identity of the authors of the papers they review.
To ensure anonymity of authorship during the review process, authors must prepare their manuscript as follows:
Authors’ names and affiliations must not appear on the title page or elsewhere in the paper. Instead, please include the number assigned to the paper by the online paper registration system under the title, for instance, by including the command inalfalse in your LaTeX source file.
Research group members, colleagues, or collaborators must not be acknowledged anywhere in the paper.
Funding sources must not be acknowledged anywhere in the submitted paper.
It is strongly suggested that the submitted file be named with the assigned submission number. For example, if your assigned paper number is 39352, name your submitted file 39352.pdf.
Source file naming must also be done with care. For example, if your name is Jane Smith and you submit a PDF file generated from a .dvi file called Jane-Smith.dvi, one can infer your authorship by looking into the PDF file.
You must also use care when referring to related past work, particularly your own. For example, avoid mentions of your previous work as “In our previous work [1,2]…” and prefer third-person referencing as “In previous work [1,2]…”. Despite the anonymity requirements, you should still include all your relevant work in the references, using the above style (omitting them could potentially reveal your identity by negation).
It is the responsibility of authors to do their best to preserve anonymity. Papers that do not follow the guidelines posted here or potentially reveal the identity of the authors are subject to immediate rejection. Having papers on arXiv is allowed per the dual submission policy outlined below.
Full papers (up to 8 pages)
Short papers (up to 4 pages)
As usual in scientific conferences, for the Oral Presentations Session, only full papers will be selected, and for the Poster/Demo Session, it will be composed of selected short papers and full papers, if necessary.
Double-blind peer review, in line with SIBGRAPI standards
Each paper will receive at least two reviews
The Program Committee is composed of experts in:
Computer vision
Medical imaging
Machine learning
Originality and relevance
Technical quality
Experimental validation
Clarity of presentation
Alignment with workshop themes
Deadline for the submission of full/short papers to the workshop – 20/Aug – 27/Aug
Notification of acceptance of the submitted abstracts/short papers – 29/Aug – 05/Set
Camera-ready due – 02/Sep – 09/Set (Hard deadline)
Workshop date (Timetable in Brazilian time) - 29/Sep (@ SIBGRAPI/2026)
Accepted papers will be published in the SIBGRAPI Workshops Proceedings (TBA), following standard indexing and formatting.
The acceptance of a work implies that at least one of its authors will register with the full registration rate at the conference and pay the publication fee for that article. If more than one author registers for the conference, the choice of who will present is up to the paper’s authors. Such information will be asked later by the program chairs. Presentations must be in English or Portuguese. Videos or remote casts are not allowed.
The program schedule will be available as soon as possible after the papers’ final acceptance decision.
A paper submitted by authors who subsequently did not present it in person at the technical meeting despite having registered at the conference will be considered a no-show paper and removed from the workshop proceedings.
By submitting papers to the SIBGRAPI 2026 Thematic Workshop, the authors acknowledge that they comply with the Code of Conduct for Authors in SBC Publications, which is available at: https://sol.sbc.org.br/index.php/indice/conduta
Prof. Flávio de Barros Vidal
University of Brasília (UnB), Brazil
Profa. Roberta Barbosa Oliveira
University of Brasília (UnB), Brazil
Prof. Angelo Amancio Duarte
State University of Feira de Santana (UEFS), Brazil
TBA.