WarpNet: Weakly Supervised Matching for Single-view Reconstruction

TL;DR

WarpNet introduces a novel deep learning method for matching objects across images in fine-grained datasets without part annotations, enabling weakly supervised single-view reconstruction with improved accuracy.

Contribution

The paper presents WarpNet, a new architecture that aligns objects across images without part annotations, leveraging dataset structure for training in an unsupervised manner.

Findings

Achieves 13.6% improvement in AP on CUB-200-2011 dataset.

Enables single-view reconstruction comparable to methods using annotated points.

Demonstrates effective matching across appearance, viewpoint, and articulation variations.

Abstract

We present an approach to matching images of objects in fine-grained datasets without using part annotations, with an application to the challenging problem of weakly supervised single-view reconstruction. This is in contrast to prior works that require part annotations, since matching objects across class and pose variations is challenging with appearance features alone. We overcome this challenge through a novel deep learning architecture, WarpNet, that aligns an object in one image with a different object in another. We exploit the structure of the fine-grained dataset to create artificial data for training this network in an unsupervised-discriminative learning approach. The output of the network acts as a spatial prior that allows generalization at test time to match real images across variations in appearance, viewpoint and articulation. On the CUB-200-2011 dataset of bird categories, we improve the AP over an appearance-only network by 13.6%. We further demonstrate that our WarpNet matches, together with the structure of fine-grained datasets, allow single-view reconstructions with quality comparable to using annotated point correspondences.