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

The success of monocular depth estimation relies on large and diversetraining sets. Due to the challenges associated with acquiring denseground-truth depth across different environments at scale, a number of datasetswith distinct characteristics and biases have emerged. We develop tools thatenable mixing multiple datasets during training, even if their annotations areincompatible. In particular, we propose a robust training objective that isinvariant to changes in depth range and scale, advocate the use of principledmulti-objective learning to combine data from different sources, and highlightthe importance of pretraining encoders on auxiliary tasks. Armed with thesetools, we experiment with five diverse training datasets, including a new,massive data source: 3D films. To demonstrate the generalization power of ourapproach we use zero-shot cross-dataset transfer}, i.e. we evaluate on datasetsthat were not seen during training. The experiments confirm that mixing datafrom complementary sources greatly improves monocular depth estimation. Ourapproach clearly outperforms competing methods across diverse datasets, settinga new state of the art for monocular depth estimation. Some results are shownin the supplementary video at https://youtu.be/D46FzVyL9I8

Towards Robust Monocular Depth Estimation: Mixing Datasets for Zero-shot Cross-dataset Transfer