notesum.ai

Published at November 3

VQ-Map: Bird's-Eye-View Map Layout Estimation in Tokenized Discrete Space via Vector Quantization

cs.CV

cs.AI

Released Date: November 3, 2024

Authors: Yiwei Zhang¹, Jin Gao¹, Fudong Ge¹, Guan Luo¹, Bing Li², Zhaoxiang Zhang³, Haibin Ling⁴, Weiming Hu⁵

Aff.: ¹State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), CASIA, School of Artificial Intelligence, University of Chinese Academy of Sciences; ²State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), CASIA, School of Artificial Intelligence, University of Chinese Academy of Sciences, People AI, Inc; ³State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), CASIA, School of Artificial Intelligence, University of Chinese Academy of Sciences, Center for Artificial Intelligence and Robotics, HKISI, CAS; ⁴Stony Brook University; ⁵State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS), CASIA, School of Artificial Intelligence, University of Chinese Academy of Sciences, School of Information Science and Technology, ShanghaiTech University

Arxiv: http://arxiv.org/abs/2411.01618v1

Methods	IoU $\uparrow$ ( $\%$ )
Methods	Drivable	Ped. Cross.	Walkway	Stopline	Carpark	Divider	Mean
OFT [36]	74.0	35.3	45.9	27.5	35.9	33.9	42.1
LSS [3]	75.4	38.8	46.3	30.3	39.1	36.5	44.4
CVT [37]	74.3	36.8	39.9	25.8	35.0	29.4	40.2
M²BEV [38]	77.2	-	-	-	-	40.5	-
BEVFusion [1]	81.7	54.8	58.4	47.4	50.7	46.4	56.6
MapPrior [17]	81.7	54.6	58.3	46.7	53.3	45.1	56.7
X-Align [34]	82.4	55.6	59.3	49.6	53.8	47.4	58.0
MetaBEV [35]	83.3	56.7	61.4	50.8	55.5	48.0	59.3
DDP [19]	83.6	58.3	61.6	52.4	51.4	49.2	59.4
VQ-Map	83.8	60.9	64.2	57.7	55.7	50.8	62.2