| 摘要 |
A computationally efficient image segmentation method is provided that processes a grayscale digital image to more clearly show textures in the underlying object shown in the digital image. A grayscale digital image is converted to an intensity matrix based on the brightness of the pixels in the image, where each matrix element represents a pixel in the digital image and has a value corresponding to the intensity, i.e., the brightness, of that pixel. The value of each matrix element is compared to the value of its nearest neighbor matrix element, and the pixel represented by the matrix element is categorized as being "dark" or "bright" based on its value, and is categorized as being "smooth" or "rough" based on the values of the nearest neighbor matrix elements. As each pixel is categorized, it is assigned a shading level corresponding to the brightness/texture matrix element value. A processed image having only the assigned shading levels is then produced, with the processed image indicating textures of the underlying object shown in the original grayscale digital image. |
