LICA Dataset

A collection of 1,183 graphic design layouts with component-level specifications and natural-language annotations, released as part of the LICA paper.

Each layout captures the full rendering spec of a design — positions, typography, images, background — along with annotations at both the layout and template level. Layouts are grouped by template (a design theme that produces multiple slide variations).

The dataset is also available on HuggingFace.

Getting started

1. Download the lica-data folder and unzip it in the repo root:

lica-dataset/
├── lica-data/
├── lica_dataset.py
├── requirements.txt
└── README.md

2. Install dependencies (Python 3.9+):

pip install -r requirements.txt

Dataset structure

lica-data/
├── metadata.csv                              # per-layout metadata
├── layouts/
│   └── <template_id>/
│       └── <layout_id>.json                  # component-level layout spec
├── images/
│   └── <template_id>/
│       └── <layout_id>.png or .mp4           # rendered layout (image or video)
└── annotations/
    ├── template_annotations.json             # template-level annotations
    └── <template_id>/
        └── <layout_id>.json                  # per-layout annotation

metadata.csv

Column	Type	Description
layout_id	string	Unique layout ID (matches filenames in layouts/, images/, annotations/)
category	string	Design category (e.g. "Presentations", "Videos", "Education", "Flyers")
template_id	string	UUID of the parent template (matches folder names)
n_template_layouts	int	Number of layouts in the template group
template_layout_index	int	Zero-based position within the template group
width	int	Canvas width in pixels
height	int	Canvas height in pixels

Layout JSON

Each file in layouts/<template_id>/<layout_id>.json contains the canvas specification and an ordered list of components:

{
  "components": [ ... ],
  "background": "rgb(252, 252, 252)",
  "width": "1920px",
  "height": "1080px",
  "duration": 3
}

Field	Type	Required	Description
components	array	yes	Ordered list of rendering components (see below)
width	string	yes	Canvas width with "px" suffix
height	string	yes	Canvas height with "px" suffix
background	string	no	CSS color for the canvas background
duration	number	no	Slide duration in seconds

Component types

Each component has a type field and CSS-like positioning/visual properties.

TEXT — positioned text element

{
  "type": "TEXT",
  "text": "Hello World",
  "left": "108px", "top": "200px", "width": "400px", "height": "50px",
  "color": "rgb(255, 255, 255)",
  "fontSize": "48px",
  "fontFamily": "League Spartan--400",
  "fontWeight": "400",
  "textAlign": "center",
  "lineHeight": "52px",
  "letterSpacing": "0em",
  "textTransform": "none",
  "fontStyle": "normal",
  "transform": "none"
}

IMAGE — positioned image

{
  "type": "IMAGE",
  "src": "https://storage.googleapis.com/lica-video/<uuid>.png",
  "left": "0px", "top": "0px", "width": "1920px", "height": "1080px",
  "transform": "none",
  "opacity": 1,
  "overflow": "hidden"
}

GROUP — container/shape element with optional clip path

{
  "type": "GROUP",
  "left": "108px", "top": "463px", "width": "555px", "height": "508px",
  "background": "rgb(255, 255, 255)",
  "backgroundColor": "rgb(255, 255, 255)",
  "clipPath": "path(\"M0,0 ...\")",
  "transform": "none"
}

Annotations

Per-layout (annotations/<template_id>/<layout_id>.json):

{
  "description": "Visual description of the specific layout",
  "aesthetics": "Notes on design style, composition, visual hierarchy",
  "tags": "comma, separated, keyword, tags",
  "user_intent": "Inferred purpose or goal of the design",
  "raw": "Concatenation of all fields above"
}

Template-level (annotations/template_annotations.json):

A JSON object keyed by template UUID. Each entry has the same fields (description, aesthetics, tags, user_intent, raw) but describes the shared design theme across all layouts in the template.

---

Quick start

from lica_dataset import LicaDataset

ds = LicaDataset("lica-data")
print(ds)
# LicaDataset(n=1183, categories=['Business Cards', 'Cards & Invitations', ...])

Filtering

Filtering methods return a new LicaDataset view and are chainable.

presentations = ds.by_category("Presentations")
template_layouts = ds.by_template("3b919d2e-539f-4b2c-8d86-7709ef65b496")
widescreen = ds.by_dimensions(1920, 1080)
portrait = ds.by_aspect_ratio("portrait")

Accessing records

layout_id = "gsessHF2ev5r4ZgwPUh5"

layout = ds.get_layout(layout_id)
annotation = ds.get_annotation(layout_id)
render_path = ds.get_render_path(layout_id)
meta = ds.get_metadata(layout_id)

Iteration

for item in ds:
    print(item["layout_id"], item["metadata"]["category"])
    # item also contains: layout, annotation, template_annotation, render_path

Convenience functions

from lica_dataset import (
    load_dataset,
    load_layouts_by_template,
    load_layouts_by_category,
    iter_template_groups,
)

ds = load_dataset("lica-data")
layouts = load_layouts_by_template("lica-data", "3b919d2e-539f-4b2c-8d86-7709ef65b496")

for template_id, group in iter_template_groups("lica-data"):
    print(template_id, len(group))

---

API reference

LicaDataset(data_root)

Method	Description
.by_category(category)	Filter by design category
.by_template(template_id)	Filter by template UUID
.by_dimensions(width, height)	Filter by exact canvas dimensions
.by_aspect_ratio(ratio)	Filter by "landscape", "portrait", or "square"
.get_layout(layout_id)	Load layout JSON
.get_annotation(layout_id)	Load per-layout annotation
.get_template_annotation(template_id)	Load template-level annotation
.get_render_path(layout_id)	Path to rendered image/video (PNG or MP4)
.get_metadata(layout_id)	Single metadata row as dict
.ids	Layout IDs in current view
.metadata	Filtered metadata as DataFrame
.categories	Sorted unique categories
.templates	Unique template IDs
.summary()	Per-category summary table

Module-level functions

Function	Description
load_dataset(data_root)	Shorthand for LicaDataset(data_root)
load_layouts_by_template(data_root, template_id)	Layout dicts for a template, sorted by index
load_layouts_by_category(data_root, category)	Layout dicts for a category
iter_template_groups(data_root)	Yields (template_id, LicaDataset) per template

Citation

@article{Hirsch2026LICA,
  title   = {LICA: Layered Image Composition Annotations for Graphic Design Research},
  author  = {Hirsch, Elad and Yadav, Shubham and Garg, Mohit and Mehta, Purvanshi},
  journal = {arXiv preprint arXiv:2603.16098},
  year    = {2026}
}

License

Creative Commons Attribution 4.0 International (CC BY 4.0).