ALIGNet: Partial-Shape Agnostic Alignment via Unsupervised Learning

TL;DR

ALIGNet is a deep learning model that aligns shapes despite missing parts by learning shape priors from datasets, using self-supervision without ground truth alignments, and promoting smooth deformations for robust, partial-shape agnostic alignment.

Contribution

This work introduces ALIGNet, a neural network that learns to align incomplete shapes by leveraging dataset priors and self-supervised training, overcoming limitations of traditional methods.

Findings

ALIGNet effectively aligns incomplete and complete shapes.

The model generalizes well to unseen data.

It produces smooth, plausible deformations even with significant shape missing parts.

Abstract

The process of aligning a pair of shapes is a fundamental operation in computer graphics. Traditional approaches rely heavily on matching corresponding points or features to guide the alignment, a paradigm that falters when significant shape portions are missing. These techniques generally do not incorporate prior knowledge about expected shape characteristics, which can help compensate for any misleading cues left by inaccuracies exhibited in the input shapes. We present an approach based on a deep neural network, leveraging shape datasets to learn a shape-aware prior for source-to-target alignment that is robust to shape incompleteness. In the absence of ground truth alignments for supervision, we train a network on the task of shape alignment using incomplete shapes generated from full shapes for self-supervision. Our network, called ALIGNet, is trained to warp complete source shapes to incomplete targets, as if the target shapes were complete, thus essentially rendering the alignment partial-shape agnostic. We aim for the network to develop specialized expertise over the common characteristics of the shapes in each dataset, thereby achieving a higher-level understanding of the expected shape space to which a local approach would be oblivious. We constrain ALIGNet through an anisotropic total variation identity regularization to promote piecewise smooth deformation fields, facilitating both partial-shape agnosticism and post-deformation applications. We demonstrate that ALIGNet learns to align geometrically distinct shapes, and is able to infer plausible mappings even when the target shape is significantly incomplete. We show that our network learns the common expected characteristics of shape collections, without over-fitting or memorization, enabling it to produce plausible deformations on unseen data during test time.