Radiation Oncology Journal
- Volume 18 Issue 4
- /
- Pages.345-354
- /
- 2000
- /
- 2234-1900(pISSN)
- /
- 2234-3156(eISSN)
Enhancement of Image Contrast in Linacgram through Image Processing
전산처리를 통한 Linacgram의 화질개선
- Suh, Hyun-Suk (Deparment of Radiation Oncology, Ewha Womans University) ;
- Shin, Hyun-Kyo (Deparment of Radiation Oncology, Ewha Womans University) ;
- Lee, Re-Na (Deparment of Radiation Oncology, Ewha Womans University)
- Published : 2000.12.01
Abstract
Purpose : Conventional radiation therapy Portal images gives low contrast images. The purpose of this study was to enhance image contrast of a linacgram by developing a low-cost image processing method. Materials and Methods : Chest linacgram was obtained by irradiating humanoid Phantom and scanned using Diagnostic-Pro scanner for image processing. Several types of scan method were used in scanning. These include optical density scan, histogram equalized scan, linear histogram based scan, linear histogram independent scan, linear optical density scan, logarithmic scan, and power square root scan. The histogram distribution of the scanned images were plotted and the ranges of the gray scale were compared among various scan types. The scanned images were then transformed to the gray window by pallette fitting method and the contrast of the reprocessed portal images were evaluated for image improvement. Portal images of patients were also taken at various anatomic sites and the images were processed by Gray Scale Expansion (GSE) method. The patient images were analyzed to examine the feasibility of using the GSE technique in clinic. Results :The histogram distribution showed that minimum and maximum gray scale ranges of 3192 and 21940 were obtained when the image was scanned using logarithmic method and square root method, respectively. Out of 256 gray scale, only 7 to 30
목적 : 방사선조사야를 확인하는 보편적인 방법인 linacgram은 저대조도(low contrast)의 영상을 보여주고 있어 정확한 영상을 확인하는데 문제점이 있다. 따라서 본 연구는 linacgram의 대조도를 높이는 저가형 확인방법을 모색하여 영상판독과 조사야 확인에 도움이 되고자 한다. 대상 및 방법: 인체모형을 사용하여 얻어진 필름 영상을 필름전용 스캐너(Diagnostic Pro)를 통해 Optical Density Scan, Histogram Equalized, Linear Histogram Based (HB), Linear Histogram Independent, Linear Optical Density (OD) Logarithmic 및 Power, Square Root scan 방식으로 디지털화 하였다. 각기 다른 방식으로 전산 입력된 영상의 신호분포도를 얻어 signal intensity를 비교한 후 pailette fitting 방식을 통해 영상을 재구성하였고 재구성된 영상을 비교 분석하였다. 실제 치료에서 얻어진 각 인체 부위별 linacgram도 동일한 방법으로 처리한 후 화질 개선도를 알아 보았다. 결과 : 인체모형을 통해 얻어진 영상의 신호 분포영역은 Logarlthmic 방식을 선택했을 때 최소값인 3192가 나왔고 Square Root방식을 사용했을 때 최대값인 21940가 나왔다. 이러한 값들을 모니터 상에서 구현할 수 있는 256 gray scale로 바꾸어 보았을 때 7