Ultrasound Image Binarization Using Basis-Based Segmentation

The patent segments the image data by transforming it into multiple data arrays classified by basis information (e.g., frequency components, spatial regions). Each segmented data array is then binarized using an adaptive threshold value specific to its characteristics, rather than applying a single global threshold to the entire image. This segmentation approach enables accurate binarization for each region while managing overall processing complexity.

The patent implements dynamic threshold values that adapt to local image characteristics. Instead of using a fixed threshold, the system calculates and applies different threshold values for different basis information categories (e.g., different frequency ranges or spatial regions). This dynamic adaptation allows the binarization process to respond to luminance variations throughout the image, improving accuracy while maintaining reasonable processing complexity through efficient algorithms.

The patent divides the image processing task into multiple segments based on basis information classification. By transforming image data into separate data arrays (e.g., by frequency, spatial region, or other characteristics), the system applies adaptive thresholding only to each segment rather than the entire image at once. This reduces the computational burden of each individual thresholding operation while collectively achieving accurate binarization across the whole image.

The patent applies different thresholding strategies and parameters to different regions or components of the image based on their specific characteristics. Each basis information category receives a tailored threshold value optimized for its local properties, rather than applying a uniform approach throughout. This local optimization improves overall binarization accuracy while avoiding the excessive computational cost of processing every pixel with the same complex adaptive algorithm.