A morphological image processing method to improve the visibility of pulmonary nodules on chest radiographic images. (March 2020)

• Record Type: Journal Article
• Authors: Kimori, Yoshitaka
• Is Part Of: Biomedical signal processing and control. Volume 57(2020)
• Publication Date: 2020-03
• Subjects: Pulmonary nodules -- Chest radiographic image -- Pulmonary nodule enhancement method -- Morphological processing; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2019.101744
• Languages: English
• ISSNs: 1746-8094

Abstract

Highlights: A new pulmonary nodule enhancement method has been described. The method is based on a morphological subtraction. The method makes it possible to selectively emphasize the nodule from other structures. The performance of the method was superior to that of conventional methods. Abstract: A pulmonary nodular shadow on a chest radiographic image is characteristic of lung cancer. However, nodular shadows often appear with very weak contrast. The target nodule shadow often overlaps with a shadow with strong contrast, such as that of the heart or rib; thus, it is difficult to emphasize the target region with ordinary contrast enhancement methods. In this article, a new pulmonary nodule enhancement method using a morphological subtraction method based on a modification of the h -dome transform is described. This method consists of two parts: a preprocessing step and nodule enhancement processing. The former reduces the uneven brightness of the background, and the latter emphasizes the nodule. With this method, the pulmonary nodule region is emphasized with the use of shape information of the nodule, whereas emphasis of the regions surrounding the nodule with different shapes is suppressed. This method also detects the boundary of the nodule as well as the edges of the surrounding anatomically normal structures. The performance of the proposed method has been verified by nodule enhancement experiments using synthetic and actual chest radiographic images from a public database. Furthermore, the efficiency of proposed method was tested by using contrast improvement index (CII), which is an index of improvement in contrast in the nodule region and its surroundings, and the results were quantitatively compared with those obtained from existing contrast emphasis methods. The proposed method has the highest value of CII, thereby confirming the effect of the contrast emphasis with this method.