AI Paper Atlas

Research map / Vision-language and multimodal / Robot action and navigation

IGen: Scalable Data Generation for Robot Learning from Open-World Images

Chenghao Gu, Haolan Kang, Junchao Lin, Jinghe Wang, Duo Wu, Shuzhao Xie, Fanding Huang, Junchen Ge, Ziyang Gong, Letian Li, Hongying Zheng, Changwei Lv and 1 more

CVPR 2026 · Vision-language and multimodal › Robot action and navigation · approach: Memory and subgoal-based manipulation · 5 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
replace Real-to-Sim digital-twin reconstruction with IGen scene reconstruction from a single imagescene reconstruction · Simpler scenes, average of 4 tasksPSNR: 17.2649 -> 27.0040Improved
replace Real-to-Sim with IGen scene reconstructionscene reconstruction · Simpler scenes, average of 4 tasksSSIM: 0.6833 -> 0.8522Improved
replace Real-to-Sim with IGen scene reconstructionscene reconstruction · Simpler scenes, average of 4 tasksLPIPS (AlexNet): 0.3235 -> 0.0630Improved
use IGen instead of Cosmos-Predict2 and TesserAct for behavior generationrobot behavior video generation · DreamGen Bench, Instruction Following and Physics Alignmentproportion of videos scored 1: nearly twice as many Instruction Following successes with Qwen-3-VL-PlusImproved
use IGen instead of TesserAct and Cosmos-Predict2 for data generationdata generation efficiency · per-sample generation under same GPU memoryefficiency: approximately 30x and 200x higher efficiencyImproved
fine-tune pi0 on IGen-generated data instead of no fine-tuningpi0 · real-world Place Bottle tasksuccess rate (%): 0.0% -> 75.0%Improved
use 100 IGen-generated samples instead of 100 teleoperated samples (same time budget)pi0 fine-tuning · real-world manipulation taskssuccess rate (%): 19.4% -> 44.5%Improved
scale IGen-generated data from 100 to 1,000 samples (ablation)pi0 fine-tuning · real-world manipulation taskssuccess rate (%): 44.5% -> 66.7%Improved
fine-tune pi0 on 1,000 IGen samples instead of teleoperation data under same collection timepi0 fine-tuning · real-world manipulation taskssuccess rate (%): 58.3% -> 66.7%Improved

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