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 bio-inspired event cameras or dynamic vision sensors are capable ofasynchronously capturing per-pixel brightness changes (called event-streams) inhigh temporal resolution and high dynamic range. However, the non-structuralspatial-temporal event-streams make it challenging for providing intuitivevisualization with rich semantic information for human vision. It calls forevents-to-video (E2V) solutions which take event-streams as input and generatehigh quality video frames for intuitive visualization. However, currentsolutions are predominantly data-driven without considering the prior knowledgeof the underlying statistics relating event-streams and video frames. It highlyrelies on the non-linearity and generalization capability of the deep neuralnetworks, thus, is struggling on reconstructing detailed textures when thescenes are complex. In this work, we propose \textbf{E2HQV}, a novel E2Vparadigm designed to produce high-quality video frames from events. Thisapproach leverages a model-aided deep learning framework, underpinned by atheory-inspired E2V model, which is meticulously derived from the fundamentalimaging principles of event cameras. To deal with the issue of state-reset inthe recurrent components of E2HQV, we also design a temporal shift embeddingmodule to further improve the quality of the video frames. Comprehensiveevaluations on the real world event camera datasets validate our approach, withE2HQV, notably outperforming state-of-the-art approaches, e.g., surpassing thesecond best by over 40\% for some evaluation metrics.

E2HQV: High-Quality Video Generation from Event Camera via Theory-Inspired Model-Aided Deep Learning