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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Depth cameras have found applications in diverse fields, such as computervision, artificial intelligence, and video gaming. However, the high latencyand low frame rate of existing commodity depth cameras impose limitations ontheir applications. We propose a fast and accurate depth map reconstructiontechnique to reduce latency and increase the frame rate in depth cameras. Ourapproach uses only a commodity depth camera and color camera in a hybrid camerasetup; our prototype is implemented using a Kinect Azure depth camera at 30 fpsand a high-speed RGB iPhone 11 Pro camera captured at 240 fps. The proposednetwork, AutoDepthNet, is an encoder-decoder model that captures frames fromthe high-speed RGB camera and combines them with previous depth frames toreconstruct a stream of high frame rate depth maps. On GPU, with a 480 x 270output resolution, our system achieves an inference time of 8 ms, enablingreal-time use at up to 200 fps with parallel processing. AutoDepthNet canestimate depth values with an average RMS error of 0.076, a 44.5% improvementcompared to an optical flow-based comparison method. Our method can alsoimprove depth map quality by estimating depth values for missing andinvalidated pixels. The proposed method can be easily applied to existing depthcameras and facilitates the use of depth cameras in applications that requirehigh-speed depth estimation. We also showcase the effectiveness of theframework in upsampling different sparse datasets e.g. video objectsegmentation. As a demonstration of our method, we integrated our frameworkinto existing body tracking systems and demonstrated the robustness of theproposed method in such applications.

AutoDepthNet: High Frame Rate Depth Map Reconstruction using Commodity Depth and RGB Cameras