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

Single image depth estimation (SIDE) plays a crucial role in 3D computervision. In this paper, we propose a two-stage robust SIDE framework that canperform blind SIDE for both indoor and outdoor scenes. At the first stage, thescene understanding module will categorize the RGB image into differentdepth-ranges. We introduce two different scene understanding modules based onscene classification and coarse depth estimation respectively. At the secondstage, SIDE networks trained by the images of specific depth-range are appliedto obtain an accurate depth map. In order to improve the accuracy, we furtherdesign a multi-task encoding-decoding SIDE network DS-SIDENet based ondepthwise separable convolutions. DS-SIDENet is optimized to minimize bothdepth classification and depth regression losses. This improves the accuracycompared to a single-task SIDE network. Experimental results demonstrate thattraining DS-SIDENet on an individual dataset such as NYU achieves competitiveperformance to the state-of-art methods with much better efficiency. Oursproposed robust SIDE framework also shows good performance for the ScanNetindoor images and KITTI outdoor images simultaneously. It achieves the topperformance compared to the Robust Vision Challenge (ROB) 2018 submissions.

Deep Robust Single Image Depth Estimation Neural Network Using Scene Understanding