影像扭轉通常用來修正影像移動或旋轉所產生之失真，因爲影像之扭轉需要很龐大的運算，故本文提出一種新的循環轉換法則(Recursive Transformation Algorithm)，以提高影像扭轉之運算速度。因所提出之轉換法則比傳統方式之速度提高10倍，故利用此種新的方法可以實現高效能之影像扭轉之運算。
Image warping is frequently used to correct the distortion caused by image translation, rotation, skewing or scaling. Since image warping involves a tremendous amount of arithmetic operations, huge computational efforts are required. This paper presents a novel recursive transformation algorithm for the second order image warping. The proposed algorithm can be employed to compute pixel coordinates efficiently by reducing the computational complexity significantly for high-speed image warping applications. The transformation algorithm is implemented by software on a computer for performance evaluation. The result of which shows a tenfold improvement in execution speed compared with that of the conventional image warping method. Using the proposed algorithm, a high-throughput image warping with efficient computational capability can be realized.
