Medical imaging pipeline

Adaptive Preprocessing for IOPA X-rays

An image preprocessing pipeline for dental X-rays that measures each image's quality and adapts its enhancement to it, instead of applying one fixed recipe.

The problem

Dental AI models depend on consistent intraoral periapical (IOPA) X-rays, but clinics use different devices and software, so brightness, contrast, sharpness and noise vary widely. Static preprocessing with fixed parameters can't cope with that spread.

Approach

• Loads DICOM, RVG and standard image formats.
• Computes per-image quality metrics for brightness, contrast, sharpness and noise.
• Chooses enhancement parameters per image from those metrics.
• Compares original, static and adaptive outputs side by side, with quantitative evaluation and plots saved for every image.

Stack

• Python
• OpenCV
• DICOM
• Image processing