AI Paper Atlas

Research map / Image generation / Autoregressive and latent generation

Learning to Infer Graphics Programs from Hand-Drawn Images

Kevin Ellis, Daniel Ritchie, Armando Solar-Lezama, Josh Tenenbaum

NeurIPS 2018 · Image generation › Autoregressive and latent generation · approach: Classical generative image synthesis · 94 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
Combine the neural proposal network with SMC rather than using either alone (ablation)Graphics spec parsing · Scenes with more objects than the training dataGeneralization to cluttered scenesImproved
Use the neural network alone with beam search instead of neural proposals plus SMC (ablation)Graphics spec parsing · Scenes with many more objects than training dataGeneralization to cluttered scenesHurt
Use SMC alone instead of neural proposals plus SMC (ablation)Graphics spec parsing · Scenes with many more objects than training dataGeneralization to cluttered scenesHurt
Replace command-by-command spec inference with a one-shot CNN-to-LSTM spec decoder (ablation)Graphics spec parsing · Cluttered scenes with many objectsParsing abilityHurt
Add a learned prior over synthesized programs to rerank candidate specsNeurally guided SMC · 100 real hand-drawn figuresTop-1 accuracy: 63% to 67%Improved
Use a learned search policy instead of sending the whole synthesis problem to SketchGraphics program synthesis · Graphics program synthesis problemsMedian search time: 274 sec -> 28 secImproved
Use a learned search policy instead of the DC component-prediction baselineGraphics program synthesis · Graphics program synthesis problemsMedian search time: 187 sec -> 28 secImproved
Regress directly from images to programs instead of factoring through specs and synthesis (ablation)Graphics program induction · Graphics program synthesis problemsTimeouts (1 hr): 2% -> 94%Hurt
Use direct image-to-program regression rather than the factored system (ablation)Graphics program induction · Graphics program synthesis problemsMedian search time: 28 sec -> 63 secHurt
Rerank candidate specs by program simplicity and structure (ablation)Neural spec predictions · Hand-drawn figuresTop-1 accuracy: 63% to 67%Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy