OPD: Single-view 3D Openable Part Detection

TL;DR

This paper introduces OPD, a method for detecting openable parts and their motion in objects from a single image, utilizing new datasets and a neural network architecture, with promising results especially on RGB images.

Contribution

The paper presents OPDRCNN, a novel neural network architecture for detecting openable parts and predicting their motion parameters from a single image, along with two new datasets for training and evaluation.

Findings

OPDRCNN outperforms baseline methods on openable part detection.

The task is challenging due to limited information in a single image.

RGB inputs yield better results than other modalities.

Abstract

We address the task of predicting what parts of an object can open and how they move when they do so. The input is a single image of an object, and as output we detect what parts of the object can open, and the motion parameters describing the articulation of each openable part. To tackle this task, we create two datasets of 3D objects: OPDSynth based on existing synthetic objects, and OPDReal based on RGBD reconstructions of real objects. We then design OPDRCNN, a neural architecture that detects openable parts and predicts their motion parameters. Our experiments show that this is a challenging task especially when considering generalization across object categories, and the limited amount of information in a single image. Our architecture outperforms baselines and prior work especially for RGB image inputs. Short video summary at https://www.youtube.com/watch?v=P85iCaD0rfc