Skip to main content

Table 2 Aggregated Jaccard Index (mean ± SD) of segmented cells in pome fruit tissues (n = 9) by deep learning models trained on data retrieved using different slice spacings and the benchmark method

From: Automatic 3D cell segmentation of fruit parenchyma tissue from X-ray micro CT images using deep learning

Model

Aggregated Jaccard Index

Apple

Pear

8

0.861 ± 0.028A

0.732 ± 0.075AB

16

0.849 ± 0.026B

0.729 ± 0.081A

32

0.846 ± 0.027C

0.719 ± 0.089BC

64

0.842 ± 0.026D

0.721 ± 0.082C

128

0.837 ± 0.026E

0.718 ± 0.083CD

256

0.836 ± 0.029E

0.718 ± 0.085C

512

0.825 ± 0.028F

0.697 ± 0.092E

Benchmark

0.730 ± 0.034G,ns

0.667 ± 0.117DE,ns

  1. Different upper case characters in the same column indicate significant differences (p < 0.05) between the segmentation results of models trained on pome fruit data retrieved using different slice spacings and the benchmark method using the watershed algorithm on different datasets on either apple or pear tissue samples. Between columns is marked with ‘ns’ if the model is not statistically different for apple or pear tissue
Back to article page

Plant Methods

ISSN: 1746-4811