Skip to content
← Back to Videos

HiP3R Reconstructs 10,000-Frame Videos in 3D

Loading video
Loading video

HiP3R aligns overlapping video chunks and refines camera poses with a hierarchical pose graph, producing globally consistent 3D reconstructions across more than 10,000 frames.

Category: research
Author: @zunnanxu
Date: 2026-10-26T00:00:00
Duration: 72.4s

As an Amazon Associate, we earn from qualifying purchases.