Skip to main content

Table 3 Accuracy of TopoRoot, DynamicRoots and DynamicRoots + for lateral root traits

From: TopoRoot: a method for computing hierarchy and fine-grained traits of maize roots from 3D imaging

Trait

 ≤ e

TopoRoot (%)

DynamicRoots (%)

DynamicRoots + (%)

1st-order lateral root count

0.04

41.4 (σ = 23.4)

75.6 (σ = 11.3)

71.5 (σ = 10.7)

0.08

49.0 (σ = 41.5)

80.7 (σ = 11.2)

71.9 (σ = 10.3)

1st-order lateral root total length

0.04

44.5 (σ = 15.0)

71.9 (σ = 13.9)

66.9 (σ = 15.4)

0.08

47.0 (σ = 15.9)

76.7 (σ = 14.9)

67.4 (σ = 14.9)

1st-order lateral root avg. length

0.04

20.4 (σ = 15.0)

38.1 (σ = 74.6)

26.5 (σ = 54.0)

0.08

25.6 (σ = 19.2)

53.2 (σ = 98.2)

26.0 (σ = 52.3)

1st-order lateral root thickness

0.04

355.0 (σ = 85.8)

424.6 (σ = 101.7)

439.2 (σ = 104.2)

0.08

342.8 (σ = 85.3)

427.7 (σ = 112.0)

450.1 (σ = 102.2)

1st-order lateral root number of children

0.04

68.3 (σ = 34.7)

152.9 (σ = 301.1)

115.6 (σ = 192.7)

0.08

80.7 (σ = 50.0)

180.9 (σ = 319.7)

115.6 (σ = 182.6)

1st-order lateral root tortuosity

0.04

4.6 (σ = 3.2)

11.9 (σ = 2.1)

12.1 (σ = 1.8)

0.08

4.5 (σ = 3.2)

10.2 (σ = 3.0)

11.3 (σ = 2.0)

1st-order lateral root emergence angle

0.04

3.4 (σ = 2.4)

N/A

N/A

0.08

3.0 (σ = 2.2)

N/A

N/A

1st-order lateral root midpoint angle

0.04

4.1 (σ = 2.7)

N/A

N/A

0.08

4.0 (σ = 2.7)

N/A

N/A

1st-order lateral root tip angle

0.04

5.4 (σ = 6.4)

N/A

N/A

0.08

5.9 (σ = 7.3)

N/A

N/A

Lateral root count

0.04

22.2 (σ = 19.5)

60.6 (σ = 13.0)

58.0 (σ = 14.8)

0.08

37.0 (σ = 56.1)

66.8 (σ = 12.4)

58.1 (σ = 14.4)

Total lateral root length

0.04

24.3 (σ = 9.3)

52.3 (σ = 19.6)

50.9 (σ = 19.0)

0.08

24.4 (σ = 11.4)

59.1 (σ = 18.8)

56.0 (σ = 17.8)

Average lateral root length

0.04

20.2 (σ = 14.8)

21.2 (σ = 22.4)

22.0 (σ = 21.6)

0.08

25.9 (σ = 19.0)

21.4 (σ = 22.9)

20.6 (σ = 20.2)

  1. Each entry gives the mean relative error and standard deviation (\(\sigma )\) for a method across all simulated models and across all noise levels up to e = 0.04 cm and e = 0.08 cm. For each trait, the method with the lowest mean error is bold-faced