Medical Image Blending With Soft Thresholds for Metal Visibility

The patent applies local quality by differentiating the blending approach for different image regions. A spatially-varying blending scheme is used where pixels are classified as either anatomy or metal objects based on intensity thresholds. Metal objects are selectively enhanced with higher opacity from the overlay image, while anatomy regions maintain the baseline image quality. This local differentiation resolves the contradiction by ensuring metal visibility without compromising overall image quality.

The patent segments the image into distinct regions based on pixel intensity characteristics. By calculating intensity differences between baseline and overlay images and applying thresholding, the system separates metal objects from anatomy. This segmentation enables independent processing of each region, allowing metal objects to be enhanced while maintaining anatomy quality, thus resolving the visibility and detection accuracy contradiction.

The patent merges the baseline image (high-quality, full-dose) with the overlay image (low-quality, low-dose) through a composite image generation process. The blending algorithm combines information from both images, using the baseline image for anatomical structure and the overlay image for metal object enhancement. This merging allows the system to achieve high image quality while using reduced radiation dose, resolving the contradiction between radiation exposure and image quality.

The patent introduces an intermediary blending algorithm that mediates between the baseline and overlay images. This intermediary process selectively transfers information from the low-quality overlay image to the composite image, enhancing metal objects while maintaining overall image quality. The intermediary blending mechanism enables the system to achieve diagnostic quality images with reduced radiation exposure.

The patent applies self-service by enabling the image processing system to automatically identify and enhance metal objects without requiring explicit segmentation algorithms. The method uses intensity-based automatic classification where pixels are assigned to metal or anatomy categories based on their intensity characteristics in the difference image. This self-service approach eliminates the need for complex manual segmentation, reducing processing complexity while maintaining metal visibility.

Instead of attempting to segment metal objects directly from the overlay image, the patent inverts the approach by working with the difference image between baseline and overlay images. By analyzing intensity differences rather than trying to isolate metal objects directly, the system achieves enhanced visibility with simpler processing. This inversion of the traditional segmentation approach reduces complexity and minimizes visual artifacts.