Table 6 Ablation results.

From: A gait recognition architecture for early screening in the assessment of Parkinson’s patients

Model

Global-info expert

Spatio-temporal expert

Local motion expert

Contour expert

Paramters count

CASIA-B

NM

BG

CL

GaitBase

7.818M

92.80

91.10

77.80

GRE-MFF-GaitBase w/o Global-Info Expert

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

7.886M

93.60

92.80

82.50

GRE-MFF-GaitBase w/o Spatio-Temporal Expert

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

7.874M

95.80

93.40

84.60

GRE-MFF-GaitBase w/o Local Motion Expert

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

7.875M

96.30

94.10

84.90

GRE-MFF-GaitBase w/o Contour Expert

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

7.881M

96.10

93.80

85.10

GRE-MFF-GaitBase (Full)

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

\(\checkmark\)

8.093M

98.70

97.50

89.20

  1. The best results of each dataset are highlighted in bold. Abbreviation explanation: NM: Normal Walking, BL: Walking with Backpack, GL: Walking in a Coat