LayerCanvas: Infographic Generation in AI Learning Assistants via Hybrid Layer-Set Diffusion

Keshav Gupta; Jeremy Owen; Christopher Lloyd; Shaunak Mishra; Manav Sharma

cs.CVOct 13, 2026

LayerCanvas: Infographic Generation in AI Learning Assistants via Hybrid Layer-Set Diffusion

Authors: Keshav Gupta, Jeremy Owen, Christopher Lloyd, Shaunak Mishra, Manav Sharma

Organizations: University of California, San Diego · LinkedIn Inc.

Abstract

Online learning platforms are increasingly deploying AI assistants to provide personalized, on-demand tutoring. These can leverage infographics to communicate complex information and improve the learning experience. However, automating effective infographic generation remains a challenging task, requiring a combination of visual richness with precise typography, structured layouts, and continued editability. Existing representation methods satisfy these goals only in part: Scalable Vector Graphics (SVG) representations preserve geometry, hierarchy, and text but provide limited aesthetic flexibility, while raster diffusion models produce richer appearances by flattening content and layout into a single image. Conditioning a diffusion model on a rasterized SVG does not resolve this conflict because the raster preserves approximate placement but discards element ownership, occlusion relationships, and the structure needed to modify and recompose the result. We introduce LayerCanvas, a hybrid approach that retains SVG as the semantic, structural, and typographic program while using diffusion to aestheticize non-text content. Rather than generating one flattened raster, LayerCanvas partitions SVG elements into z-ordered sets of mutually non-overlapping elements and processes them through communicating parallel RGB generation streams. The resulting layers are independently editable and deterministically recomposed, while text is rendered from preserved vector metadata. Our approach combines rich visual synthesis with accurate typography and localized editing. LayerCanvas substantially improves text fidelity over state-of-the-art raster image generators while achieving the highest overall aesthetic quality. We further demonstrate non-destructive content editing, localization, asset reuse, and animation.

Figures & tables

Figure 1: LayerCanvas generates editable, text-faithful infographics. Current SOTA text-to-image models such as Qwen-Image-2512 ( Qwen Team, 2025a ) and Ideogram-4 ( Ideogram AI, 2026 ) flatten the graphic and often garble text (top). Given a vector layout, LayerCanvas generates ordered RGB layers and composes them with exact vector text (middle), enabling localization and illustration edits without changing unrelated content (bottom).

Figure 2: LayerCanvas layer generation. Top: Parallel denoising produces aligned non-text layers, which are composed in paint order before exact SVG text is restored. Bottom: Layer-set attention produces residuals for shared frozen MMDiT updates, and the solver advances every layer.

	Text fidelity			VLM judge (0-5)
Method	Char	Exact	Recall	Appeal	Typog.	Cov.	Avg
SVG (LLM-authored)	0.856	0.676	0.964	2.37	4.02	4.90	3.76
Qwen-Image-2512 ( Qwen Team, 2025a )	0.080	0.002	0.149	2.69	1.00	1.45	1.71
Ideogram-4 ( Ideogram AI, 2026 )	0.657	0.360	0.744	2.92	2.29	3.89	3.03
LayerCanvas (Ours)	0.822	0.623	0.942	2.99	3.81	4.75	3.85

Table 1: Primary comparison on test set. Char and Exact measure placement-exact OCR, while Recall measures text recovered anywhere in the image. Appeal, Typog., and Cov. are VLM scores from 0-5, with Avg denoting their mean. Best synthesis result in each column is bold .

Figure 3: Qualitative four-way comparison Columns show, left to right, the SVG control, two prompt-only SOTA diffusion baselines (Qwen-Image-2512 Qwen Team (2025a) and Ideogram-4 Ideogram AI (2026) ), and LayerCanvas. The corresponding next row shows the zoomed in regions for the highlighted text regions. Prompt-only models produce appealing raster illustrations but garble or invent text and drift from the prescribed layout whereas LayerCanvas keeps the layer layout and composites exact, legible vector text.

Figure 4: Representation ablation: Each column shows the source SVG, fixed-noise base generation, and the translated version by perturbing the base SVG for LayerCanvas and the single-raster as control. Red arrows show the shift from the base position in the translated cases. LayerCanvas better preserves object identity and unrelated layout content.

Method	Char	Exact	Recall	Appeal	Typog.	Cov.	Avg
Single-raster Conditioning	0.851	0.568	0.783	2.78	3.56	4.32	3.55
LayerCanvas	0.823	0.530	0.795	2.84	3.76	4.72	3.77

Table 2: Comparison of Different Conditioning Representations on validation set

Figure 5: Independent animation of one LayerCanvas region. A Text2Video-Zero-inspired image-diffusion trajectory produces an open-to-closed blink. Only the selected eye region changes across the five frames; the surrounding illustration, layout, and vector typography remain fixed.

Figure 6: Closed-loop SVG correction. Each row shows an incorrect SVG and render, the direct SVG edit, and the corrected render. The edited element moves to the requested location while unrelated content remains stable.

Appendix figures & tables13 assets

Supplementary material from the paper’s appendix.

Appendix

Figure 7: The paired training record. Each column is one artifact the pipeline emits for the same sample. Left: the structural input, an SVG with color, fonts, and textures stripped and text set in a neutral face, which supplies geometry, hierarchy, and paint order without leaking target appearance. Middle: the aestheticized, text-free target that the diffusion streams are trained to produce; the editable-visual regions have been restyled into rich illustrations while every text region is left blank. Right: the final paired target after the preserved vector text layer is composited on top. Diffusion is never asked to render glyphs.

Figure 8: Representative dataset infographics. The corpus spans diverse layout archetypes (above-below iceberg, circular cycle, hero-plus-steps, comparison split, guided procedure, staged timeline), both light and dark themes, and a range of illustration styles, while preserving crisp typography and clean editable-visual footprints in every sample. Text is always rendered from vector metadata, never by the diffusion editor.

Figure 9: Cross-element color structure ( $n=1{,}000$ ). (a) Per-channel Pearson correlation in CIELAB between mean illustration color and mean palette color; coupling is strongest on the $b^{\ast}$ (blue-yellow) axis. (b) Observed illustration $\leftrightarrow$ palette coupling exceeds a shuffled null ( $p_{\mathrm{perm}}<10^{-3}$ ). (c) Joint-structure and learnability summaries: canonical correlation $\rho_{1}$ , variance explained by the first two joint principal components versus a null, and the $R^{2}$ gain on $b^{\ast}$ from conditioning on the palette. Together these show the dataset induces coordinated, non-random color composition across layers.

	Text fidelity			VLM judge (0-5)
Method	Char	Exact	Recall	Appeal	Typog.	Cov.	Craft	Avg
SVG (LLM-authored) †	0.856	0.676	0.964	2.37	4.02	4.90	3.91	3.76
Qwen-Image-2512	0.080	0.002	0.149	2.69	1.00	1.45	1.51	1.71
Ideogram-4	0.657	0.360	0.744	2.92	2.29	3.89	2.49	3.03
LayerCanvas (Ours)	0.822	0.623	0.942	2.99	3.81	4.75	3.59	3.85

Appendix

Table 3: Full evaluation, including the craft judge criterion omitted from Table 1 . Text fidelity is measured at

768{\times}512

and includes placement-exact ground-truth-region OCR (char / exact) and placement-agnostic full-image word recall ( Recall ); judge criteria are

0

5

, higher better; Avg is the mean of appeal, typography, and coverage. † LLM-authored SVG recolor reference.

	$768{\times}512$ OCR			4096-pixel OCR
Method	Char	Exact	Recall	Char	Exact	Recall
SVG (LLM-authored) †	0.856	0.676	0.964	0.883	0.595	0.814
Qwen-Image-2512	0.080	0.002	0.149	0.077	0.002	0.150
Ideogram-4	0.657	0.360	0.744	0.643	0.268	0.727
LayerCanvas (Ours)	0.822	0.623	0.942	0.853	0.543	0.806

Appendix

Table 4: OCR resolution sensitivity. Placement-exact character accuracy ( Char ) and exact-match rate ( Exact ), together with placement-agnostic word recall ( Recall ).

Judge	SVG	Qwen	Ideogram	LayerCanvas
Claude Opus 4.8 Anthropic (2026)	3.76	1.71	3.03	3.85
Gemini 3.7 Flash Google DeepMind (2026)	3.93	1.99	3.54	3.99
GPT-5.6 Sol OpenAI (2026)	3.81	1.99	3.32	3.81

Appendix

Table 5: Multi-judge robustness at

768{\times}512

. Each entry is the mean of illustration appeal, typography quality, and semantic coverage. LayerCanvas is the highest-scoring learned synthesis method under every judge.

Figure 10: Additional qualitative comparisons. Columns show the LLM-authored SVG reference, Qwen-Image-2512, Ideogram-4, and LayerCanvas, respectively.

Figure 11: Additional qualitative comparisons. Columns show the LLM-authored SVG reference, Qwen-Image-2512, Ideogram-4, and LayerCanvas, respectively.

Figure 12: Diffusion inference compute. Generation time and peak allocated GPU memory on one H200 increase with the number of grouped layers. Each point is the mean of two 20-step generations; error bars show the measured minimum and maximum.

Figure 13: Editable text representation and localization. A held-out LayerCanvas result is generated without glyphs, while the source SVG text nodes remain a separate editable vector layer. Their composition produces the final document. The lower row localizes the same unchanged generated image layers into French, Arabic, and Japanese by re-rendering only the text layer.

Figure 14: Asset reuse. Three generated layers are taken from one document and arranged into a new composition. Each layer carries its own alpha, allowing independent repositioning, scaling, restyling, and reordering.

Figure 15: Layout retargeting. The same generated module illustration, code panels, and explanatory cards are rearranged for slide, poster, and mobile canvases. Image groups preserve their aspect ratios, while text remains separately renderable for each target layout.

Figure 16: Command-driven editing. Three natural-language commands are converted to targeted remove, swap, and add operations. Red boxes indicate the edited regions; unselected layers are reused unchanged.

Vector Graphics Generation

See all 26 papers

Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs
Aug 26, 2026Matthew Perlman, James Beetham, Niels Da Vitoria Lobo +2Vector Graphics Generation VLM Adaptation
University of Massachusetts Amherst · University of Central Florida
CreatiParser: Generative Image Parsing of Raster Graphic Designs into Editable Layers
Apr 21, 2026Weidong Chen, Dexiang Hong, Zhendong Mao +4
School of Information Science and Technology, University of Science and Technology of China, Hefei 230027, China · School of Information Science and Technology, University of Science and Technology of China, and are also with the Institute of Artificial Intelligence, Hefei Comprehensive National Science Center, Hefei 230027, China · Individual Researcher +1
DrawAI: Agentic Benchmark and Workflow for Making Raster Images Editable
Aug 1, 2026Pu Cao, Qingye Kong, Xuedan Yin +5Image Reconstruction
Beijing University of Posts and Telecommunications · Tsinghua University · Beijing Institute of Technology

Aug 26, 2026J/K move · Enter open · S save

Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs

Matthew Perlman, James Beetham, Niels Da Vitoria Lobo, Amrit Singh Bedi+1

University of Massachusetts Amherst · University of Central Florida

Vector Graphics Generation VLM Adaptation Iterative Refinement VLM Evaluation Constrained Generative Modeling

Open paper AskPDF arXiv

Scalable Vector Graphics (SVGs) power much of the modern visual ecosystem, yet state-of-the-art generative models focus almost entirely on rasterized images. We explore whether inference-time methods can unlock SVG generation capabilities in off-the-shelf vision-language models (VLMs). We systematically evaluate a constrained iterative refinement harness that combines visual feedback, structured editing, and constrained decoding to characterize the capabilities and limitations of current VLMs for SVG generation. Across multiple VLMs and generation settings, we find that constrained decoding improves compilation success rates, while iterative refinement reveals a deficit in visual reasoning and self-correction. Our results highlight both the promise and current limitations of using inference-time methods to adapt general-purpose VLMs for SVG generation.

Abstract

Vector Graphics Generation

Explore similar work

Evaluating Constrained Iterative Refinement for Scalable Vector Graphics Generation with Off-the-Shelf VLMs

CreatiParser: Generative Image Parsing of Raster Graphic Designs into Editable Layers

DrawAI: Agentic Benchmark and Workflow for Making Raster Images Editable