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

There have been two streams in the 3D detection from point clouds:single-stage methods and two-stage methods. While the former is morecomputationally efficient, the latter usually provides better detectionaccuracy. By carefully examining the two-stage approaches, we have found thatif appropriately designed, the first stage can produce accurate box regression.In this scenario, the second stage mainly rescores the boxes such that theboxes with better localization get selected. From this observation, we havedevised a single-stage anchor-free network that can fulfill these requirements.This network, named AFDetV2, extends the previous work by incorporating aself-calibrated convolution block in the backbone, a keypoint auxiliarysupervision, and an IoU prediction branch in the multi-task head. As a result,the detection accuracy is drastically boosted in the single-stage. To evaluateour approach, we have conducted extensive experiments on the Waymo Open Datasetand the nuScenes Dataset. We have observed that our AFDetV2 achieves thestate-of-the-art results on these two datasets, superior to all the prior arts,including both the single-stage and the two-stage 3D detectors. AFDetV2 won the1st place in the Real-Time 3D Detection of the Waymo Open Dataset Challenge2021. In addition, a variant of our model AFDetV2-Base was entitled the "MostEfficient Model" by the Challenge Sponsor, showing a superior computationalefficiency. To demonstrate the generality of this single-stage method, we havealso applied it to the first stage of the two-stage networks. Withoutexception, the results show that with the strengthened backbone and therescoring approach, the second stage refinement is no longer needed.

AFDetV2: Rethinking the Necessity of the Second Stage for Object Detection from Point Clouds