Image Processing
Image Processing
This domain tests latent-image formation, film processing, and digital image handling. Questions cover density and contrast, characteristic curves, exposure index, windowing, DICOM and PACS.
Core principles
- Light or X-rays expose silver halide grains, forming a latent image. Developer reduces exposed grains to black silver; fixer removes unexposed halide and hardens the emulsion.
- Film density = log10(I0/It). The characteristic curve has toe, straight-line and shoulder regions; contrast is the slope.
- Digital detectors separate acquisition from display. The histogram represents pixel values; window width sets contrast and window level sets brightness.
- Exposure index (EI) indicates detector signal, not patient dose. Compare EI with the manufacturer's target and use deviation index where available.
- DICOM stores images and metadata; PACS archives and distributes them. Lossless compression preserves diagnosis.
Worked example
A digital chest image looks bright but the EI is above the target range. Windowing can improve display, but the acquisition was overexposed. The correct reasoning is: adjust display for review, check EI and clinical indication, and repeat only if diagnostic information is compromised.
Traps
- Equating digital brightness with exposure.
- Repeating solely because EI is high when the image is diagnostic.
- Confusing window width (contrast) with window level (brightness).
- Assuming PACS compression is always lossless.