Skip to main content

Table 1 Statistics of samples, volume (presented by the LIDC-IDRI dataset), and accuracy (with 95% confidence intervals) of classification using the two types of images on the training, validation, and test datasets

From: Decoding intra-tumoral spatial heterogeneity on radiological images using the Hilbert curve

Ā 

Training

Validation

Test

Dataset

Ā Benign

428

52

52

Ā Malignant

321

40

40

Ā Volume (avg)

1221.05

985.33

1069.61

Accuracy (95% CI)

Ā Original CT images

95.69% (94.29ā€“97.09%)

75.05% (66.21ā€“83.89%)

73.22% (64.18ā€“82.26%)

Ā Hilbert matrix images

93.45% (91.86ā€“95.04%)

86.36% (69.35ā€“93.37%)

85.54% (78.35ā€“92.73%)

  1. CI confidence intervals