Reduction of Artifacts in Degraded CT Image by Adaptive Wiener Filter

• Published in: Iyo denshi to seitai kogaku Vol. 40; no. 1; pp. 1 - 6
• Main Authors: FUNAMA, Yoshinori, NOGUCHI, Yoshio, SHIMAMURA, Masamichi
• Format: Journal Article
• Language: Japanese
• Published: Japanese Society for Medical and Biological Engineering 2002; 一般社団法人 日本生体医工学会
• Subjects: Computerized tomography; Medical imaging; Pulmonary diseases; Signal to noise ratio; Spurious signal noise
• ISSN: 0021-3292; 2185-5498
• DOI: 10.11239/jsmbe1963.40.1

In the screening for lung cancer by helical computed tomography (CT), the detection of Type A according to Noguchi's classification is important. However, Noguchi's Type A, which shows low CT values, is difficult to detect. In particular, at low doses the detection rate further decreases because of artifact noise in the upper lung fields and in the dorsal areas of the middle lung. To reduce artifact noise, we processed the images containing artifact noise (degraded images), which were obtained by using a phantom with a CT value similar to that of Noguchi's Type A, with an adaptive Wiener filter. Moreover, clinical images diagnosed as primary lung cancer were processed with the same method for the phantom. The reduction in artifact noise in such images was assessed by signal-to-noise ratio (SNR) improvement and by visual evaluation. The detection of Noguchi's Type A with low CT value has been improved by the use of an adaptive Wiener filter.