English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We present DistillFlow, a knowledge distillation approach to learning opticalflow. DistillFlow trains multiple teacher models and a student model, wherechallenging transformations are applied to the input of the student model togenerate hallucinated occlusions as well as less confident predictions. Then, aself-supervised learning framework is constructed: confident predictions fromteacher models are served as annotations to guide the student model to learnoptical flow for those less confident predictions. The self-supervised learningframework enables us to effectively learn optical flow from unlabeled data, notonly for non-occluded pixels, but also for occluded pixels. DistillFlowachieves state-of-the-art unsupervised learning performance on both KITTI andSintel datasets. Our self-supervised pre-trained model also provides anexcellent initialization for supervised fine-tuning, suggesting an alternatetraining paradigm in contrast to current supervised learning methods thathighly rely on pre-training on synthetic data. At the time of writing, ourfine-tuned models ranked 1st among all monocular methods on the KITTI 2015benchmark, and outperform all published methods on the Sintel Final benchmark.More importantly, we demonstrate the generalization capability of DistillFlowin three aspects: framework generalization, correspondence generalization andcross-dataset generalization.

Learning by Distillation: A Self-Supervised Learning Framework for Optical Flow Estimation