moment invariants
时间: 2024-09-16 16:06:22 浏览: 36
Visual Pattern Recognition by Moment Invariants
"Moments invariants"通常是指在统计学和计算机视觉领域中的一种概念,它涉及到图像特征描述。Moments are quantities derived from the intensity or color distribution of an image region, such as pixels, and they are usually represented by a set of coefficients called central moments or raw moments. Central moments are calculated relative to the image's center of mass, while raw moments are not.
Central moments (also known as Hu moments) are invariant under translation, rotation, and scaling transformations, making them useful for feature detection and recognition tasks because these properties help ensure that features extracted from transformed images will match those from the original image. By using invariants like moments, algorithms can be more robust to variations in pose and lighting conditions.
To compute moments, you typically sum up products of pixel intensities with increasing powers of distance from the center point. For example, first-order moments give information about the centroid, second-order moments provide information on size and orientation, and higher-order moments contribute to more complex shape characteristics.
阅读全文