| 摘要: | 一般言之，目標物影像之顯示與否均有其特定之複雜性存在，其邊綠之成長亦有其邏輯性之變動。以往在計算機科學應用及人工智慧處理上，將目標物之影像均以整體形態加以分析處理，但本文是在可行條件下，以簡易表達方法一一柯氏複性程序法，將機率複性變爲定值長度，並利用對應法則及杜林機完成辨認工作，其優點除使演算時間減少外，更能增大辨認正確率。在此使用結法，可以直接表達出柯氏之使用結果。
爲了使物體的方位標準化且提高辨認率；本文採用最小串列辨認圖形法，可分爲二種：1環層串列法，或稱爲非線性串結排列。2直線串列法，或稱爲線性串法排列。兩者同時匹配使用時可節省很多時間，環層串列法把物體方位標準化，而直線串列則應用於增加辨認速率的工作。相關事項：標型辨認，串結排列編碼法，環層編碼，影像處理，柯氏複雜性程序，杜林機、自動機，形態無關法則，推入機、金字塔法。
This paper is concerned with the intrinsic complexity of image patterns. Previous works on the decomposition of pictures into their basic building blocks form the basis of this paper. However, the work presented in this paper will look at the simplest description over all others. The complexity of Kolmogorov, given the domain, is the minimum length of program so that a Turing Machine can compute functions in finite time. The Tolmogorov complexity program is performed so that each 2-D image pattern can be decomposed into many rows. Each row is identical to a scanning line of a digital TV camera, If a binary pixel occurs in a given cell, it is coded as ”0” or ”1”. When the cells in a row are constant, the occurrence of ”0” or ”1”, which represents the object pixel or background respectively, forms a string array. In this manner, the complexity of high level image pattern can be converted into a low level binary system for 2-D or 3-D pattern recognition. Therefore, the string array approach is a good implementation of Kolmogorov process.
In order to normalize the object's orientation and to speed up the recognition, a Minimum Array for Recognition Chart (MAR Chart) is created. There are two kinds of MAR arrays obtained from the chart: 1. Circular Layer String Array (CLSA), or called Nonlinear String Array (NSA), and 2. Straight Line String Array (SLSA), or called Linear String Array (LSA). A lot of time can be saved in matching if both of NSA and LSA are used together. The CLSA is used for normalization of object's orientation, while the SLSA is applied for increasing the speed of recognition.
Index Terms--Pattern Recognition, String Array Code, Circular Layer Code, Image Processing, Kolmogorov Complexity Program, Turing Machine, Automaton, Context-free Grammar, Pushdown Machine, Pyramid Approach. |
