Never Look Back: Understanding Persistence in 3D Object Memory from Egocentric Videos
arXiv cs.AI · 8 Oct 2026, 03:59 AEST
A new study explores how an embodied assistant can create a persistent 3D object memory from egocentric videos. The system, called Ledger, improves object localization and memory retention significantly.
Key facts
- We present Ledger, a persistent 3D object memory that combines object locations, their histories, and contextual descriptions.
- It associates observations across the recording and retains objects after they leave the view, including those the person never touches.
- Our memory raises HD-EPIC accuracy from 29.7% to 42.6%, UCS-Bench accuracy from 33.8% to 38.5% and localizes Ego4D objects with a 0.99 m median error on returned predictions.
- Our analyses identify complementary roles for temporal persistence, contextual descriptions, and retrieval.
Read aloud
A new study explores how an embodied assistant can create a persistent 3D object memory from egocentric videos. The system, called Ledger, improves object localization and memory retention significantly. We present Ledger, a persistent 3D object memory that combines object locations, their histories, and contextual descriptions. It associates observations across the recording and retains objects after they leave the view, including those the person never touches. Our memory raises HD-EPIC accuracy from twenty nine point seven percent to forty two point six percent, UCS-Bench accuracy from thirty three point eight percent to thirty eight point five percent and localizes Ego4D objects with a zero point nine nine m median error on returned predictions. Our analyses identify complementary roles for temporal persistence, contextual descriptions, and retrieval.
Sources in this cluster
- Never Look Back: Understanding Persistence in 3D Object Memory from Egocentric Videos
arXiv cs.AI · 8 Oct 2026, 03:59 AEST