English

Instance-wise Depth and Motion Learning from Monocular Videos

Computer Vision and Pattern Recognition 2020-04-09 v2 Machine Learning Robotics

Abstract

We present an end-to-end joint training framework that explicitly models 6-DoF motion of multiple dynamic objects, ego-motion and depth in a monocular camera setup without supervision. Our technical contributions are three-fold. First, we propose a differentiable forward rigid projection module that plays a key role in our instance-wise depth and motion learning. Second, we design an instance-wise photometric and geometric consistency loss that effectively decomposes background and moving object regions. Lastly, we introduce a new auto-annotation scheme to produce video instance segmentation maps that will be utilized as input to our training pipeline. These proposed elements are validated in a detailed ablation study. Through extensive experiments conducted on the KITTI dataset, our framework is shown to outperform the state-of-the-art depth and motion estimation methods. Our code and dataset will be available at https://github.com/SeokjuLee/Insta-DM.

Keywords

Cite

@article{arxiv.1912.09351,
  title  = {Instance-wise Depth and Motion Learning from Monocular Videos},
  author = {Seokju Lee and Sunghoon Im and Stephen Lin and In So Kweon},
  journal= {arXiv preprint arXiv:1912.09351},
  year   = {2020}
}

Comments

Project page at https://sites.google.com/site/seokjucv/home/instadm

R2 v1 2026-06-23T12:51:22.620Z