Table 5 Processing memory (RAM) and total prediction time.

From: A dataset for environmental sound recognition in embedded systems for autonomous vehicles

Classifier

Notebook

Notebook

Raspberry Pi

Raspberry Pi

RAM

(Std)

Total pred. time

(Std)

RAM

(Std)

Total pred. time

(Std)

(MB)

(ms)

(MB)

(ms)

SVM

34.1

(4.7)

223.4

(11.4)

32.4

(8.2)

764.4

(39.6)

LR

1.1

(0.1)

220.7

(11.8)

0.2

(0.03)

749.2

(14.6)

RF

401.1

(65.4)

227.2

(12.9)

456.0

(11.3)

1,020.9

(31.1)

ANN

20.6

(9.2)

250.1

(16.0)

0.1

(0.02)

752.5

(12.5)

CNN 1D

7.5

(0.5)

252.1

(20.0)

4.1

(1.9)

750.0

(13.6)

CNN 2D

11.7

(0.2)

30.2

(7.2)

19.4

(2.5)

47.6

(22.0)

  1. Comparison between the predictive algorithms in the notebook and Raspberry Pi.