Skip to main content

Table 2 Comparisons of recognition accuracies on Sneeze-Cough dataset

From: Recognizing flu-like symptoms from videos

Brief description

Prec. %

Rec. %

Acc. %

cuboid + BoW + χ2 (baseline)

49.5

45.0

30.9

cuboid + AMK I U c S=6

69.5

47.5

39.3

cuboid + AMK I U b S=27

78.7

47.2

43.4

cuboid + AMK I B S=2

50.0

54.3

35.2

cuboid + AMK II

55.3

62.1

41.3

HOGHOF + BoW + χ2 (baseline)

52.6

50.0

34.4

HOGHOF + AMK I U c S=3

63.7

48.3

37.9

HOGHOF + AMK I U b S=27

69.7

40.8

34.6

HOGHOF + AMK I B S=5

60.8

49.2

37.3

HOGHOF + AMK II

58.9

64.4

44.4

  1. Comparisons of recognition accuracies on Sneeze-Cough dataset. Here shorthands of AMK I and II are used for AMK type I and type II kernels, respectively. U c (U b ) refers to the concentric (block) partition scheme in unary extension. B is for binary extension. Note a different accuracy measure of TP TP + FP + FN is used here.