Skip to main content

Table 1 Details of images used for training. For water point an additional set was added due to a lack of images containing pipes and taps

From: Exploring convolutional neural networks and spatial video for on-the-ground mapping in informal settlements

Category

Total images

Total objects

Total images(II)

Total objects(II)

Drain

98

98

–

–

Trash

67

84

–

–

M.Water

74

86

–

–

W.Bucket

49

96

–

–

Tire

55

88

–

–

W.Point

59

61

94

104