Lung Fields Segmentation Algorithm in Chest Radiography - Advances in Image and Graphics Technologies - page 143

Table 2. Segmentation Result under Different Conditions

Without feature
images

With feature With feature images and the

lung outline correction

Condition

images

Overlap (%) 82.18

89.07 90.26

section, we will show the segmentation performance under three conditions: without
feature images, with feature images, with feature images and lung outline correction.
The segmentation results were given in Table . Fig.1 shows the segmentation per-

formance of lung outline under the three conditions. As can be seen from the Fig.1,
the lung segmentation algorithm based on feature images and lung outline correction
proposed in this paper can obtain the optimal segmentation results.

Ⅱ

(a)

(b)

(c)

Fig. 1. The Segmentation results without feature images and with feature images are shown in
(a) and (b), respectively. The final segmentation result with feature images and lung outline
correction is shown in (c).