Patch2CAD: Patchwise Embedding Learning for In-the-Wild Shape Retrieval from a Single Image

TL;DR

This paper introduces Patch2CAD, a novel method that learns patchwise embeddings to connect 2D image patches with 3D CAD models, enabling robust shape retrieval and pose estimation from single images in real-world scenes.

Contribution

It proposes a patchwise embedding learning approach that improves 3D shape retrieval and pose estimation by establishing correspondences between image patches and CAD model patches.

Findings

Outperforms state-of-the-art methods on ScanNet in real-world scenarios.

Enables robust shape retrieval without requiring exact CAD matches.

Improves 3D pose estimation accuracy from single images.

Abstract

3D perception of object shapes from RGB image input is fundamental towards semantic scene understanding, grounding image-based perception in our spatially 3-dimensional real-world environments. To achieve a mapping between image views of objects and 3D shapes, we leverage CAD model priors from existing large-scale databases, and propose a novel approach towards constructing a joint embedding space between 2D images and 3D CAD models in a patch-wise fashion -- establishing correspondences between patches of an image view of an object and patches of CAD geometry. This enables part similarity reasoning for retrieving similar CADs to a new image view without exact matches in the database. Our patch embedding provides more robust CAD retrieval for shape estimation in our end-to-end estimation of CAD model shape and pose for detected objects in a single input image. Experiments on in-the-wild, complex imagery from ScanNet show that our approach is more robust than state of the art in real-world scenarios without any exact CAD matches.