Table 1 The summary of the results of the introduced framework.

From: Computer vision and statistical insights into cycling near miss dynamics

Models

Factors

APa

Statistical weightb

Norm. weightc

Model1-NightNet

Nightime

0.885

0.268

0.101

Daytime

0.885

0.120

0.045

Dawn_dusk

0.885

0.417

0.157

Model2-GlareNet

Glare

0.883

0.099

0.037

Model3-PrecipitationNet(b)

Clear

0.959

0.117

0.044

Rain

0.959

0.281

0.106

Snow

0.959

0.199

0.075

Model4-FogNet

Fog

0.862

0.241

0.091

Model5-SlipNet

Wet surface

0.918

0.241

0.091

Model6-Object_detection

Person

0.75

0.074

0.028

Bicycle

0.79

0.201

0.076

Car

0.81

0.089

0.034

Bus

0.77

0.100

0.038

Motorbike

0.81

0.368

0.139

Truck

0.77

0.034

0.013

Model7-Cyclinglane

Cyclinglane

0.91

0.1235

0.047

  1. aThe average precision calculated for each model on the test sets.
  2. bThe absolute value of the statistical weight of the second logistic regression model computed based on the coefficient (B) statistics.
  3. cThe normalised version of the statistical weight is introduced in the previous column.