目的：應用軟片於劑量驗証，在軟片數位化之過程會因使用儀器的光源、機械動作影響，造成數位化影像間存在差異。先前研究曾利用ROC(receiver operator characteristics)方式比較兩種不同光源掃描器得到沒明顯差異之結果；然而，上述方式依賴視力來判別影像差異之方法並不適用於劑量驗証之需求。我們評估兩台不同型態之掃描器，應用於軟片劑量驗証的特性；此二台掃描器屬於同一家醫院所擁有，購買於相同年份（1998年），並利用作為日常機器品質保證工具之用。材料與方法：利用複製片(Kodak，RA Duplicating Film)複製標準光密度片(Kodak photographic step tablet no.3，21 steps，O. D. range 0.05~3.05)及其他實驗用之軟片，將測試片分別經過兩種不同掃描器數位化，一台為螢光燈光源(Vidar VXR-12 PLUS)，另一台掃描器為雷射光源 (Lumiscan-50)。實驗用之軟片，設計用來比較並量化兩者之暖機效應，訊號及雜訊比，百分標準誤差，光密度範圍，穿透性假影，及干涉性假影(Moiré artifact)等特性，上述特性將造成像數值(pixel value)的變動，因而造成劑量轉換時的誤差。結果：掃描器的訊號對雜訊比、百分標準誤差及光源穿透能力，會根據不同光密度值變化而影響，雷射光源掃描器在上述情況有較穩定的表現；另一方面，螢光光源掃描器則產生較少的穿透性假影，且不會有干涉性假影產生。結論：應用掃描器在軟片劑量驗証需考慮其使用目的，螢光光源掃描器能提供較高的空間解析度，適用於光照野及輻射照野一致性或星狀照射(star shoot)等檢查，而利用作為劑量檢查時，雷射光源掃描器則提供相對較寬的劑量分析範圍及較低的雜訊產生。

Purpose: In radiographic dosimetry, the performance of the digitizing system is highly dependent on the illumination mechanism or characteristics of light source. Previous analysis of the two types of digitizers by ROC (receiver operator characteristics) method has shown no significant differences. However, ROC method by visual determination is not sensitive for dosimetry purpose. We evaluated and compared the distinctive features between two digitizers that were purchased in the same particular year (1998's) by the same hospital. They have been routinely used for QA tools and film dosimetry. Materials and Methods: The duplicate films (Kodak, RA Duplicating Film) were made to be replicas of standard O.D. strip (Kodak photographic step tablet no.3, 21 steps, O.D. range 0.05～3.05) and other test films. Two digitizers were used to digitize those films. One digitizer was the Vidar VXR-12 PLUS with fluorescent light source and the other was Lumiscan-50 with laser beam source. Test films were made in order to investigate their warm-up effect, signal-to-noise ratio, percentage standard deviation, optical density range, transmission artifact and Moiré artifacts. Those features might be affected when films were converted into corresponding value through the above-mentioned digitizers. Results: Varying in O.D. range influence the feature of signal-to-noise ratio, percentage standard deviation, and transmitting capacity that performed by digitizers. The digitizer with a laser beam source was shown more stable in such circumstances. On the other hand, the digitizer with a broadband fluorescent light source had a minor transmission artifact and no Moiré artifacts than the previous one. Conclusions: Selection of a digitizer for film dosimetry depends on what kind of data one would want to acquire. A fluorescence light digitizer has a higher spatial resolution and is more suitable for radiation field-light field coincident and star shoot check. On the other hand, a laser beam digitizer offers wilder dose range and less noise.