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 ability of accurate depth prediction by a convolutional neural network(CNN) is a major challenge for its wide use in practical visual simultaneouslocalization and mapping (SLAM) applications, such as enhanced camera trackingand dense mapping. This paper is set out to answer the following question: Canwe tune a depth prediction CNN with the help of a visual SLAM algorithm even ifthe CNN is not trained for the current operating environment in order tobenefit the SLAM performance? To this end, we propose a novel online adaptationframework consisting of two complementary processes: a SLAM algorithm that isused to generate keyframes to fine-tune the depth prediction and anotheralgorithm that uses the online adapted depth to improve map quality. Once thepotential noisy map points are removed, we perform global photometric bundleadjustment (BA) to improve the overall SLAM performance. Experimental resultson both benchmark datasets and a real robot in our own experimentalenvironments show that our proposed method improves the overall SLAM accuracy.While regularization has been shown to be effective in multi-taskclassification problems, we present experimental results and an ablation studyto show the effectiveness of regularization in preventing catastrophicforgetting in the online adaptation of depth prediction, a single-taskregression problem. In addition, we compare our online adaptation frameworkagainst the state-of-the-art pre-trained depth prediction CNNs to show that ouronline adapted depth prediction CNN outperforms the depth prediction CNNs thathave been trained on a large collection of datasets.

Online Mutual Adaptation of Deep Depth Prediction and Visual SLAM