Unsupervised Domain Adaptive Object Detection using Forward-Backward Cyclic Adaptation
Preprint
Abstract
We present a novel approach to perform the unsupervised domain adaptation for object detection through forward-backward cyclic (FBC) training. Recent adversarial training based domain adaptation methods have shown their effectiveness on minimizing domain discrepancy via marginal feature distributions alignment. However, aligning the marginal feature distributions does not guarantee the alignment of class conditional distributions. This limitation is more evident when adapting object detectors as the domain discrepancy is larger compared to the image classification task, e.g. various number of objects exist in one image and the majority of content in an image is the background. This motivates us to learn domain invariance for category level semantics via gradient alignment. Intuitively, if the gradients of two domains point in similar directions, then the learning of one domain can improve that of another domain. To achieve gradient alignment, we propose Forward-Backward Cyclic Adaptation, which iteratively computes adaptation from source to target via backward hopping and from target to source via forward passing. In addition, we align low-level features for adapting holistic color/texture via adversarial training. However, the detector performs well on both domains is not ideal for target domain. As such, in each cycle, domain diversity is enforced by maximum entropy regularization on the source domain to penalize confident source-specific learning and minimum entropy regularization on target domain to intrigue target-specific learning. Theoretical analysis of the training process is provided, and extensive experiments on challenging cross-domain object detection datasets have shown the superiority of our approach over the state-of-the-art.Keywords
All Related Versions
This publication has 20 references indexed in Scilit:
- Domain Adaptive Faster R-CNN for Object Detection in the Wild2018
- Cross-Domain Weakly-Supervised Object Detection Through Progressive Domain Adaptation2018
- Maximum Classifier Discrepancy for Unsupervised Domain Adaptation2018
- Semantic Foggy Scene Understanding with Synthetic DataInternational Journal of Computer Vision, 2018
- Multi-view 3D Object Detection Network for Autonomous Driving2017
- Adversarial Discriminative Domain Adaptation2017
- Deep Residual Learning for Image Recognition2016
- The Cityscapes Dataset for Semantic Urban Scene Understanding2016
- You Only Look Once: Unified, Real-Time Object Detection2016
- A theory of learning from different domainsMachine Learning, 2009