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

Holistic understanding and reasoning in 3D scenes play a vital role in thesuccess of autonomous driving systems. The evolution of 3D semantic occupancyprediction as a pretraining task for autonomous driving and robotic downstreamtasks captures finer 3D details compared to methods like 3D detection. Existingapproaches predominantly focus on spatial cues, often overlooking temporalcues. Query-based methods tend to converge on computationally intensive Voxelrepresentation for encoding 3D scene information. This study introducesS2TPVFormer, an extension of TPVFormer, utilizing a spatiotemporal transformerarchitecture for coherent 3D semantic occupancy prediction. Emphasizing theimportance of spatiotemporal cues in 3D scene perception, particularly in 3Dsemantic occupancy prediction, our work explores the less-explored realm oftemporal cues. Leveraging Tri-Perspective View (TPV) representation, ourspatiotemporal encoder generates temporally rich embeddings, improvingprediction coherence while maintaining computational efficiency. To achievethis, we propose a novel Temporal Cross-View Hybrid Attention (TCVHA)mechanism, facilitating effective spatiotemporal information exchange acrossTPV views. Experimental evaluations on the nuScenes dataset demonstrate asubstantial 3.1% improvement in mean Intersection over Union (mIoU) for 3DSemantic Occupancy compared to TPVFormer, confirming the effectiveness of theproposed S2TPVFormer in enhancing 3D scene perception.

S2TPVFormer: Spatio-Temporal Tri-Perspective View for temporally coherent 3D Semantic Occupancy Prediction