AdaStereo: A Simple and Efficient Approach for Adaptive Stereo Matching
arXiv:2004.04627
Abstract
Recently, records on stereo matching benchmarks are constantly broken by end-to-end disparity networks. However, the domain adaptation ability of these deep models is quite poor. Addressing such problem, we present a novel domain-adaptive pipeline called AdaStereo that aims to align multi-level representations for deep stereo matching networks. Compared to previous methods for adaptive stereo matching, our AdaStereo realizes a more standard, complete and effective domain adaptation pipeline. Firstly, we propose a non-adversarial progressive color transfer algorithm for input image-level alignment. Secondly, we design an efficient parameter-free cost normalization layer for internal feature-level alignment. Lastly, a highly related auxiliary task, self-supervised occlusion-aware reconstruction is presented to narrow down the gaps in output space. Our AdaStereo models achieve state-of-the-art cross-domain performance on multiple stereo benchmarks, including KITTI, Middlebury, ETH3D, and DrivingStereo, even outperforming disparity networks finetuned with target-domain ground-truths.
arXiv commentAccepted by CVPR 2021
References in corpus (23)
- Deep Residual Learning for Image Recognition
- Instance Normalization: The Missing Ingredient for Fast Stylization
- Fully Convolutional Networks for Semantic Segmentation
- Learning Transferable Features with Deep Adaptation Networks
- Unsupervised Representation Learning by Predicting Image Rotations
- ADVENT: Adversarial Entropy Minimization for Domain Adaptation in Semantic Segmentation
- Conditional Adversarial Domain Adaptation
- The Cityscapes Dataset for Semantic Urban Scene Understanding
- Computing the Stereo Matching Cost with a Convolutional Neural Network
- FCNs in the Wild: Pixel-level Adversarial and Constraint-based Adaptation
- Curriculum Domain Adaptation for Semantic Segmentation of Urban Scenes
- Revisiting Batch Normalization For Practical Domain Adaptation
- End-to-End Learning of Geometry and Context for Deep Stereo Regression
- Central Moment Discrepancy (CMD) for Domain-Invariant Representation Learning
- Unsupervised Domain Adaptation through Self-Supervision
- Self-supervised Domain Adaptation for Computer Vision Tasks
- Batch-Instance Normalization for Adaptively Style-Invariant Neural Networks
- Cascade Cost Volume for High-Resolution Multi-View Stereo and Stereo Matching
- Stereo DSO: Large-Scale Direct Sparse Visual Odometry with Stereo Cameras
- Learning for Disparity Estimation through Feature Constancy
- Channel Normalization in Convolutional Neural Network avoids Vanishing Gradients
- Stereo Vision-based Semantic 3D Object and Ego-motion Tracking for Autonomous Driving
- Domain-invariant Stereo Matching Networks