Evaluating Deep Clustering Algorithms on Non-Categorical 3D CAD Models

TL;DR

This paper benchmarks deep clustering algorithms on large-scale non-categorical 3D CAD models, introducing a new evaluation workflow and an ensemble-based comparison method to address unique challenges in this domain.

Contribution

It presents the first benchmarking framework for deep clustering on 3D CAD models, including a novel ensemble-based evaluation approach for non-categorical data.

Findings

Seven baseline deep clustering methods evaluated.

Identified fundamental challenges in clustering non-categorical 3D data.

Proposed an ensemble-based clustering comparison approach.

Abstract

We introduce the first work on benchmarking and evaluating deep clustering algorithms on large-scale non-categorical 3D CAD models. We first propose a workflow to allow expert mechanical engineers to efficiently annotate 252,648 carefully sampled pairwise CAD model similarities, from a subset of the ABC dataset with 22,968 shapes. Using seven baseline deep clustering methods, we then investigate the fundamental challenges of evaluating clustering methods for non-categorical data. Based on these challenges, we propose a novel and viable ensemble-based clustering comparison approach. This work is the first to directly target the underexplored area of deep clustering algorithms for 3D shapes, and we believe it will be an important building block to analyze and utilize the massive 3D shape collections that are starting to appear in deep geometric computing.