notesum.ai

Published at November 20

A Resource Efficient Fusion Network for Object Detection in Bird's-Eye View using Camera and Raw Radar Data

cs.CV

cs.AI

Released Date: November 20, 2024

Authors: Kavin Chandrasekaran, Sorin Grigorescu, Gijs Dubbelman, Pavol Jancura

Arxiv: http://arxiv.org/abs/2411.13311v1

Accuracy Metrics

Computational Metrics

Methods

Modality

AP(%)

\uparrow

AR(%)

\uparrow

F1(%)

\uparrow

RE(m)

\downarrow

AE(^∘)

\downarrow

\downarrow

Avg

FPS

\uparrow

\sigma

\downarrow

Model

size

\downarrow

GPU

cost

\downarrow

FFTRadNet† [16]

93.45

83.35

88.11

0.12

0.15

3.23

68.46

2.19

39.2

2.01

TFFTRadNet† [27]

ADC

90.80

88.31

89.54

0.15

0.13

9.08

54.37

4.28

109.5

2.04

ADCNet [28]

ADC

91.9

0.13

0.11

CMS [39]

RD&C

96.9

83.49

89.69

0.45

7.7

ROFusion [40]

RD&RPC&C

91.13

95.29

93.16

0.13

0.21

3.33

56.11

1.55

87.2

2.87

EchoFusion [41]

RT&C

96.95

93.43

95.16

0.12

0.18

25.61

102.5

Ours

RD&C

95.75

91.35

93.49

0.11

0.09

6.58

58.91

1.28

79.8

2.06