Paper Overview
- Field: Computer Vision
- Authors: Pradhaan S Bhat, Naveen Chandra R, Rishubh Parihar
- Published: 2026-06-20
- arXiv: 2506.16438
Summary
Text prompts and 2D conditioning interfaces offer weak, ambiguous control over spatial transformations in image editing, particularly for large object motions and camera changes. Prior work has used 3D primitives (such as boxes), but only as loose conditional signals indicating approximate object placement, rather than as specifications of the desired transformation.
This paper instead treats 3D boxes as a structured specification: the user provides input and output boxes for an edit, turning image editing into a well-defined geometric problem. In this box-based thinking interface, each box face is color-coded to convey 3D orientation, enabling precise control over translation, rotation, scaling, and viewpoint changes in real images — while preserving scene and object identity and recovering previously unseen object regions.
To anchor transformations in the scene's appearance, the authors introduce a depth-aligned planar ground serving as a global reference frame, colorized with depth-aware cues. With this structural conditioning, an image generator produces consistent results even under large transformations.
The system is trained in two stages — on synthetic multi-object scenes and a small number of real-world videos from Objectron — yet generalizes to complex in-the-wild real images. The method operates directly on real photographs and substantially outperforms recent state-of-the-art approaches on large-scale 3D editing.
---
*Auto-collected on 2026-06-21.*