From: A multisource fusion framework driven by user-defined knowledge for egocentric activity recognition
Proposed method
Method in [15]
Method in [16] (%)
\( \mathcal{M} \)-20 (Nc = 8) (%)
\( {\mathcal{M}}_S \)-6 (Nc = 4) (%)
Maximum pooling (%)
Vision
53.2
75.8
75.0
78.4
Sensors
65.8
72.2
49.5
69.0
Fusion
72.3
85.2
80.5
83.7