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

Recent works have shown the superior robustness of four-dimensional (4D)Radar-based three-dimensional (3D) object detection in adverse weatherconditions. However, processing 4D Radar data remains a challenge due to thelarge data size, which require substantial amount of memory for computing andstorage. In previous work, an online density reduction is performed on the 4DRadar Tensor (4DRT) to reduce the data size, in which the density reductionlevel is chosen arbitrarily. However, the impact of density reduction on thedetection performance and memory consumption remains largely unknown. In thispaper, we aim to address this issue by conducting extensive hyperparamtertuning on the density reduction level. Experimental results show thatincreasing the density level from 0.01% to 50% of the original 4DRT densitylevel proportionally improves the detection performance, at a cost of memoryconsumption. However, when the density level is increased beyond 5%, only thememory consumption increases, while the detection performance oscillates belowthe peak point. In addition to the optimized density hyperparameter, we alsointroduce 4D Sparse Radar Tensor (4DSRT), a new representation for 4D Radardata with offline density reduction, leading to a significantly reduced rawdata size. An optimized development kit for training the neural networks isalso provided, which along with the utilization of 4DSRT, improves trainingspeed by a factor of 17.1 compared to the state-of-the-art 4DRT-based neuralnetworks. All codes are available at: https://github.com/kaist-avelab/K-Radar.

Enhanced K-Radar: Optimal Density Reduction to Improve Detection Performance and Accessibility of 4D Radar Tensor-based Object Detection