AI Paper Atlas

Research map / Image generation / Text-to-image diffusion

LEGO: Learning EGOcentric Action Frame Generation via Visual Instruction Tuning

Bolin Lai, Xiaoliang Dai, Lawrence Chen, Guan Pang, James M Rehg, Miao Liu

ECCV 2024 · Image generation › Text-to-image diffusion · approach: Layout-to-image control · 19 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
enrich action labels with GPT-curated, instruction-tuned VLLM descriptions (ablation)LEGO · Ego4DEgoVLP score: 62.19 -> 62.91Improved
add VLLM image embedding as extra conditioning (ablation)LEGO · Ego4DEgoVLP score: 62.91 -> 65.35Improved
add VLLM text embedding as extra conditioning (ablation)LEGO · Ego4DEgoVLP score: 62.91 -> 63.29Improved
use both image and text embeddings jointly (ablation)LEGO · Ego4DEgoVLP score: 65.35 -> 65.65Improved
finetune VLLM via visual instruction tuning instead of using off-the-shelf embeddings (ablation)LEGO · Ego4DEgoVLP score: 64.57 -> 65.65Improved
finetune VLLM for description generation versus off-the-shelf VLLM (ablation)prompt enrichment · user study, description alignment% aligned with frames: 27% -> 87% (Ego4D); 30% -> 84% (Epic-Kitchens)Improved
LEGO versus IP2P on FIDLEGO · Epic-KitchensFID: 20.64 -> 21.57Hurt
scale up training with both Ego4D and Epic-KitchensLEGO · Epic-KitchensFID: 21.57 -> 19.64Improved
scale up training with both datasetsLEGO · Ego4DFID: 23.83 -> 23.56No clear effect
LEGO versus IP2P on image-to-text metricsLEGO · Ego4DBLIP image-to-text similarity: +0.38% and +0.14%Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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