Self-Supervised Object Detection via Generative Image Synthesis

TL;DR

This paper introduces SSOD, a novel end-to-end self-supervised object detection framework that uses controllable GANs for image synthesis, significantly improving detection accuracy without requiring bounding box annotations or 3D assets.

Contribution

The paper presents the first end-to-end analysis-by-synthesis framework with controllable GANs for self-supervised object detection, enhancing detection performance and domain adaptation.

Findings

Outperforms prior state-of-the-art self-supervised methods on KITTI and Cityscapes.

Surpasses rendering-based methods without needing 3D CAD assets.

Demonstrates effective domain adaptation without labels.

Abstract

We present SSOD, the first end-to-end analysis-by synthesis framework with controllable GANs for the task of self-supervised object detection. We use collections of real world images without bounding box annotations to learn to synthesize and detect objects. We leverage controllable GANs to synthesize images with pre-defined object properties and use them to train object detectors. We propose a tight end-to-end coupling of the synthesis and detection networks to optimally train our system. Finally, we also propose a method to optimally adapt SSOD to an intended target data without requiring labels for it. For the task of car detection, on the challenging KITTI and Cityscapes datasets, we show that SSOD outperforms the prior state-of-the-art purely image-based self-supervised object detection method Wetectron. Even without requiring any 3D CAD assets, it also surpasses the state-of-the-art rendering based method Meta-Sim2. Our work advances the field of self-supervised object detection by introducing a successful new paradigm of using controllable GAN-based image synthesis for it and by significantly improving the baseline accuracy of the task. We open-source our code at https://github.com/NVlabs/SSOD.