cs.RONov 13, 2026

When Does Cross-View Consistency Matter in Multi-View Appearance Augmentation? A Controlled Study for Visuomotor Policies

Authors: Zixuan Jiang

Organizations: College of Engineering, Peking University

Abstract

Multi-view visual augmentation often seeks consistency across cameras, but it is unclear which forms of consistency matter for downstream control. We use simulation-state replay to manipulate cross-view coupling while preserving the per-camera perturbation distributions, physical trajectories, and action labels. Across seven LIBERO manipulation tasks, marginal-matched changes in global appearance, ghost injection, and target color have small average costs for a late-fusion Diffusion Policy: point estimates are near zero or negative, and the largest upper endpoint of the 90% posterior credible intervals is 2.55 percentage points. In contrast, geometrically anti-coupled observations reduce success by 3.30 percentage points relative to coupled observations (90% CrI [0.64, 6.28]), with displacement magnitude, direction marginals, and the number of affected views matched. Camera interventions reveal substantial dependence on wrist-view geometry, while feature analysis shows appearance sensitivity patterns associated with the factors randomized during training. ACT shows a higher target-color cost estimate than DP, with uncertainty that includes zero and sensitivity to task composition. Within the tested tasks and policy configurations, these findings motivate selective consistency constraints: appearance coupling can sometimes be relaxed, whereas task-relevant geometry and cross-view spatial correspondence warrant protection.

Figures & tables

Appendix figures & tables29 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Pose6DAug: Physically Plausible Multi-view Object Swapping for Robot Data Augmentation

    Jun 18, 2026Jonghoon Lee, Seong Hyeon Park, Byungwoo Jeon +2Multi-View ConsistencyData Augmentation

  2. One Demo is Worth a Thousand Trajectories: Action-View Augmentation for Visuomotor Policies

    Jun 17, 2026Chuer Pan, Litian Liang, Dominik Bauer +4Data AugmentationGeneralization in Robotic Manipulation

  3. InfiNoVA: Infinite Novel View Augmentation for Viewpoint Invariant Robot Policies

    Sep 23, 2026Sai Puneeth Reddy Gottam, Elmar Rueckert, Vedant DaveRobot LearningRobustness of VLA Models