PosterSentry is a lightweight, CPU-optimized multimodal classifier that determines whether a PDF is a scientific poster or a non-poster (paper, proceedings, newsletter, abstract book, etc.).

Part of the quality control pipeline for posters.science, a platform for making scientific conference posters Findable, Accessible, Interoperable, and Reusable (FAIR).

Developed by the FAIR Data Innovations Hub at the California Medical Innovations Institute (CalMI²).

Version

Version	Date	Notes
1.1.0	2026-08-28	Feature extraction moved off the AGPL-licensed PyMuPDF to pdfplumber (text and structure) and pypdfium2 (page rendering), both permissively licensed, so the whole stack is MIT-compatible. The head was retrained on the re-extracted features; held-out accuracy is 89.2%.
1.0.0	2026-08-18	Head trained on the human-validated corpus: 3,381 documents labeled by a three-reviewer survey (Krippendorff's alpha 0.79) with blinded adjudication of the 439 contested documents.

Earlier unversioned weights (April 2026) were trained on heuristically labeled data and are superseded; they remain available in the repository history.

Related Models & Tools

Resource	Description	Link
PosterSentry	Multimodal poster classifier (this model)	fairdataihub/poster-sentry
poster-sentry	Installable classifier package	GitHub
poster-sentry-training	Training code and replication	GitHub
poster-sentry-training-data	Human-validated training dataset (3,381 samples)	HuggingFace
poster-sentry-evaluation-paper-code	Reproducible analysis for the paper	GitHub
Llama-3.1-8B-Poster-Extraction	Poster → structured JSON extraction	fairdataihub/Llama-3.1-8B-Poster-Extraction
poster2json	Python library for poster extraction	PyPI · Docs · GitHub
poster-json-schema	DataCite-based poster metadata schema	GitHub
Platform	posters.science	posters.science

Pipeline Position

PosterSentry sits at the front of the posters.science pipeline. It screens incoming PDFs before the expensive Llama-based extraction:

PDF Input
   │
   ▼
┌──────────────┐     ┌───────────────────────────────────┐     ┌──────────────┐
│ PosterSentry │ ──► │ Llama-3.1-8B-Poster-Extraction    │ ──► │ poster2json  │
│ (classify)   │     │ (extract structured metadata)      │     │ (validate)   │
└──────────────┘     └───────────────────────────────────┘     └──────────────┘
   poster? ✓              raw text → JSON schema                  FAIR output

Architecture

Three feature channels concatenated into a 542-dimensional vector, fed to a single LogisticRegression:

Channel	Features	Dimension	Signal
Text	model2vec (potion-base-32M) embedding	512	Semantic content
Visual	Color stats, edge density, FFT spatial complexity, whitespace	15	Visual layout
Structural	Page count, area, font diversity, text blocks, density	15	PDF geometry

Each classifier head is a single linear layer stored as a numpy .npz file (10 KB). Inference is pure numpy, no torch required at prediction time.

Performance

Evaluated against human-validated labels (three-reviewer survey with blinded adjudication of contested documents):

Metric	Value
Held-out accuracy (508 documents)	89.2% (95% CI 86.2 to 91.6)
Macro F1 (held-out)	0.892
Precision / Recall / F1 (poster)	0.887 / 0.897 / 0.892
Precision / Recall / F1 (non-poster)	0.897 / 0.886 / 0.892
Out-of-fold accuracy (all 3,381 documents, 5-fold)	88.0%
Classifier speed	under one second per file on a CPU (PDF parsing is the bottleneck)

Errors concentrate where the human panel itself divided: out-of-fold agreement is 90.3% on documents the panel rated unanimously and 73.2% on documents decided two to one.

Top Features by Importance

Standardized logistic regression coefficients of the trained head (positive pushes toward poster):

Rank	Feature	Coefficient	Signal
1	`line_count`	+2.80	Posters pack many short text lines
2	`edge_density`	+2.61	Visually busy layouts
3	`page_width_pt`	+2.22	Posters are physically wide
4	`img_width`	+2.16	Large rendered width
5	`size_per_page_kb`	+1.82	Dense, high-resolution single pages
6	`color_diversity`	+1.70	Posters are visually rich
7	`page_count`	-1.64	More pages pushes away from poster
8	`font_count`	+1.52	Many fonts across poster panels
9	`is_landscape`	+1.26	Many posters are landscape
10	`file_size_kb`	-1.25	Multi-page documents are bigger overall

Structural and visual cues carry most of the decision; three embedding dimensions appear among the fifteen largest coefficients.

Training Data

Trained on 3,381 documents with human-validated labels, zero synthetic data:

Class	Count	Label provenance
Poster	1,686	Three-reviewer survey; unanimous panel label or blinded adjudication
Non-poster	1,695	Three-reviewer survey; unanimous panel label or blinded adjudication

Three reviewers independently rated all 3,570 candidate documents (inter-rater Krippendorff's alpha 0.79); the 439 documents without a unanimous panel were settled in a blinded adjudication review. After removing 182 near-duplicate documents and 7 with unavailable PDFs, the remaining 3,381 form the training corpus. Applied to the full corpus of 30,195 readable repository PDFs labeled as posters, PosterSentry classifies 77.2% as posters: more than one in five records labeled as posters is something else.

Training data: fairdataihub/poster-sentry-training-data

Usage

Python API

python

from poster_sentry import PosterSentry

sentry = PosterSentry()
sentry.initialize()

# Classify a PDF (uses text + visual + structural features)
result = sentry.classify("document.pdf")
print(f"Is poster: {result['is_poster']}, Confidence: {result['confidence']:.2f}")
# {'is_poster': True, 'confidence': 0.97, 'path': 'document.pdf'}

# Batch classification
results = sentry.classify_batch(["poster1.pdf", "paper.pdf", "newsletter.pdf"])

Installation

bash

pip install git+https://github.com/fairdataihub/poster-repo-qc.git

# Or install from source
git clone https://github.com/fairdataihub/poster-repo-qc.git
cd poster-repo-qc
pip install -e ".[train]"

Training

python scripts/train_poster_sentry.py --n-per-class 2000

Training completes in ~40 minutes on CPU (PDF rendering is the bottleneck, not the classifier).

Model Specifications

Attribute	Value
Embedding backbone	minishlab/potion-base-32M (model2vec StaticModel)
Embedding dimension	512
Visual features	15 (color, edge, FFT, whitespace)
Structural features	15 (page geometry, fonts, text blocks)
Total input dimension	542
Classifier	LogisticRegression (sklearn) + StandardScaler
Head file size	10 KB (.npz)
Precision	float32
GPU required	No (CPU-only)
License	MIT

System Requirements

CPU: Any modern CPU (no GPU needed)
RAM: ≥4GB
Python: ≥3.10
Dependencies: numpy, model2vec, scikit-learn, pdfplumber, pypdfium2, Pillow

Citation

bibtex

@software{poster_sentry_2026,
  title = {PosterSentry: Multimodal Scientific Poster Classifier},
  author = {O'Neill, Jamey and Portillo, Dorian and Zeinali, Nahid and Soundarajan, Sanjay and Blake, Gerard and Sarin, Parth and Buttrick, Adam and Patel, Bhavesh},
  year = {2026},
  version = {1.1.0},
  url = {https://huggingface.co/fairdataihub/poster-sentry},
  note = {Part of the posters.science initiative}
}

License

This model is released under the MIT License.

Acknowledgments

FAIR Data Innovations Hub at California Medical Innovations Institute (CalMI²)
posters.science platform
MinishLab for the model2vec embedding backbone
HuggingFace for model hosting infrastructure
Funded by The Navigation Fund (10.71707/rk36-9x79), "Poster Sharing and Discovery Made Easy"