We study the implementation of a meat-quality grading system, using the concept of the "marbling score", as well as image processing, neural network techniques and multiple regression analysis. The marbling score is a measure of the distribution density of fat in the rib-eye region. We identify five features used for grading meat images. For the evaluation of the five features, we propose a method of image binarization using a three-layer neural network developed on the basis of inputs given by a professional grader and a system of meat-quality grading based on the evaluation of three of five features with multiple regression analysis. Experimental results show that the system is effective.

We developed a system for automatically sketching patterns on the surface of porcelain. The system consists of two parts-measurement for the 3D porcelain piece using slit lighting and automatic pattern sketching. Since we classified porcelain pieces into two types by shape, two 3D measurements were developed. We propose calculating of 3D coordinates simply and quickly. Sketching uses numerical control. Numerical control data for sketching patterns onto porcelain were generated using results measured by the measurement and CAD systems. Experimental results verified the system's effectiveness.

In answer to what represents the intrinsic information-processing capability of the pattern classification system, Cover [1] has defined the separating capacity, and has derived it for the linear machine and the so-called ¿ machine. In this paper, the separating capacity of a multithreshold classification element is obtained. It is shown that the capacity of a multithreshold threshold element with k thresholds-k-threshold element-in n-dimensional space is 2(n + k). A linear machine is a special case in the k-threshold element with k = 1; therefore, its capacity becomes 2(n + 1) from the above result. Further, although it is intuitively apparent that the larger the number of thresholds, the more powerful the information-processing capability of the k-threshold element, using the capacity as a measure of this capability, we may now state that the separating power of the k-threshold element increases linearly with respect to k.

The two-layer threshold-element networks with feed-forward connections can be classified into two types, according to whether the inputs are fed directly to the second layer (Type II) or not (Type I). It is shown that there always exists a Type I network (with an appropriate number of elements in the first layer) which realizes the same discriminant function as that of any given Type II network; furthermore, parameters of the Type I network are determined only by that of the Type II network.