Application of phase congruency for discriminating some lung diseases using chest radiograph

Abstract

A novel procedure using phase congruency is proposed for discriminating some lung disease using chest radiograph. Phase congruency provides information about transitions between adjacent pixels. Abrupt changes of phase congruency values between pixels may suggest a possible boundary or another feature that may be used for discrimination. This property of phase congruency may have potential for deciding between disease present and disease absent where the regions of infection on the images have no obvious shape, size, or configuration. Five texture measures calculated from phase congruency and Gabor were shown to be normally distributed. This gave good indicators of discrimination errors in the form of the probability of Type I Error (δ) and the probability of Type II Error (β). However, since 1 - δ is the true positive fraction (TPF) and β is the false positive fraction (FPF), an ROC analysis was used to decide on the choice of texture measures. Given that features are normally distributed, for the discrimination between disease present and disease absent, energy, contrast, and homogeneity from phase congruency gave better results compared to those using Gabor. Similarly, for the more difficult problem of discriminating lobar pneumonia and lung cancer, entropy and homogeneity from phase congruency gave better results relative to Gabor.

Figure 1

Digital chest X-rays of (a) normal lung, (b) LC in left upper zone, and (c) PNEU in right and left lower zone (source: Insitute of Respiratory Medicine, Kuala Lumpur).

Figure 2

Chest radiograph normalization process following method in [16].

Figure 3

Images of normalization procedure.

Figure 4

Varying of wavelet scale numbers, n.

Figure 5

Varying o with n = 5.

Figure 6

Varying λ _min⁡ with n = 5, o = 6.

Figure 7

Varying α and σ with n = 5, o = 6, and λ _min⁡ = 8.

Figure 8

Varying d with n = 5, o = 6, λ _min⁡ = 8, α = 3, and σ = 0.55.

Figure 9

Varying γ with n = 5, o = 6, λ _min⁡ = 8, α = 3, σ = 0.55, and d = 1.

Figure 10

Varying c with n = 5, o = 6, λ _min⁡ = 8, α = 3, σ = 0.55, d = 1, and γ = 50.

Figure 11

Varying k with n = 5, o = 6, λ _min⁡ = 8, α = 3, σ = 0.55, d = 1, γ = 50, and c = 0.2.

Figure 12

QQ plots for (a) homogeneity PC texture measures for NL, LC, and PNEU, respectively, and (b) homogeneity Gabor texture measures for NL, LC, and PNEU, respectively.

Figure 13

The ROC curves for discriminating: (a), (b) LC and NL, (c), (d) PNEU and NL, and (e), (f) LC and PNEU using PC and Gabor features, respectively.