SuperDepth: Self-Supervised, Super-Resolved Monocular Depth Estimation

By Sudeep Pillai, Rares Ambrus, Adrien Gaidon. 2018

229 citations recorded by the source corpus.

Recent techniques in self-supervised monocular depth estimation are approaching the performance of supervised methods, but operate in low resolution only. We show that high resolution is key towards high-fidelity self-supervised monocular depth prediction. Inspired by recent deep learning methods for Single-Image Super-Resolution, we propose a subpixel convolutional layer extension for depth super-resolution that accurately synthesizes high-resolution disparities from their corresponding low-resolution convolutional features. In addition, we introduce a differentiable flip-augmentation layer that accurately fuses predictions from the image and its horizontally flipped version, reducing the effect of left and right shadow regions generated in the disparity map due to occlusions. Both contributions provide significant performance gains over the state-of-the-art in self-supervised depth and pose estimation on the public KITTI benchmark. A video of our approach can be found at https://youtu.be/jKNgBeBMx0I.

Subfields: Computer Vision and Pattern Recognition, Media Technology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Is Pseudo-Lidar needed for Monocular 3D Object detection? (2021, background): “It is itself a thriving research area: the community has pushed toward accurate dense depth prediction via supervised [35, 29, 19, 48, 26, 18, 32, 13, 31] as well as self-supervised [46, 72, 60, 16, 15, 62] methods.”
Disentangling Monocular 3D Object Detection (2019, background): “An expansion stage of ROI-10D [19] takes advantage of depth information provided by SuperDepth [22], which itself is learned in a self-supervised manner.”
Learning Representations by Predicting Bags of Visual Words (2020, methodology): “The information to be reconstructed can be provided by a different view [21, 51, 75] or sensor [16].”
Digging Into Self-Supervised Monocular Depth Estimation (2018, methodology): “978 SuperDepth + pp [47] (1024 × 382) S 0.”
On the Uncertainty of Self-Supervised Monocular Depth Estimation (2020, methodology): “Other methods improved efficiency [53, 50] to enable deployment on embedded devices, or accuracy by simulating a trinocular setup [56], jointly learning for semantic [79], using higher resolution [51], GANs [1], sparse inputs from visual odometry [2] or a teacher-student scheme [52].”
HR-Depth: High Resolution Self-Supervised Monocular Depth Estimation (2020, methodology): “Some methods also conduct high-resolution experiments where they train their models with larger images (Godard et al. 2019; Pillai et al. 2019).”
3D Packing for Self-Supervised Monocular Depth Estimation (2019, background): “Recent works [38, 51] in self-supervised monocular depth estimation use this concept to super-resolve estimates and further improve performance.”
Self-Supervised Monocular Depth Hints (2019, methodology): “Our non pretrained model beats SuperDepth [26] in 6 out of 7 metrics (tied in 1), and compares favourably to the concurrent work monoResMatch [30], which makes use of a significantly more complex network compared to our encoderdecoder architecture.”
Online Bag-of-Visual-Words Generation for Unsupervised Representation Learning (2020, background): “Many of them are formulated as reconstruction problems [1, 30, 46, 55, 57, 71, 81, 82, 84], where the reconstruction target is defined at image pixel level.”