Skip to main content

Table 2 Performance comparisons between the HOCG descriptors and its based regional features

From: Abnormal event detection in crowded scenes using histogram of oriented contextual gradient descriptor

Approaches

AUC(%)

Improvement(%)

Dimension

Gray

63.20 71.94

8.74

64

HOCG

8

SF [42]

71.48 75.74

4.26

30

HOCG

8

GCM [34]

75.32 76.44

1.12

48

HOCG

8

3D gradient

62.62 74.04

11.42

64

HOCG

8