Objects Can Move: 3D Change Detection by Geometric Transformation Constistency

TL;DR

This paper introduces a novel 3D change detection method for AR/VR and robotics that identifies moved objects solely based on scene changes, without prior object assumptions, using depth map differences and geometric consistency.

Contribution

It presents a new scene change-based object discovery approach that detects and segments moving objects through depth differences and graph cut optimization, outperforming existing methods.

Findings

Achieves state-of-the-art performance on 3RScan dataset.

Does not require prior object models or assumptions.

Effectively segments moving objects based on rigid motion detection.

Abstract

AR/VR applications and robots need to know when the scene has changed. An example is when objects are moved, added, or removed from the scene. We propose a 3D object discovery method that is based only on scene changes. Our method does not need to encode any assumptions about what is an object, but rather discovers objects by exploiting their coherent move. Changes are initially detected as differences in the depth maps and segmented as objects if they undergo rigid motions. A graph cut optimization propagates the changing labels to geometrically consistent regions. Experiments show that our method achieves state-of-the-art performance on the 3RScan dataset against competitive baselines. The source code of our method can be found at https://github.com/katadam/ObjectsCanMove.