In this paper, a new method of composing a multiclass classifier using pairwise classifiers is proposed. A “Resemblance Model ” is exploited to calculate a posteriori probability for combining pairwise classifiers. We proved the validity of this model by using approximation of a posteriori probability formula. Using this theory, we can obtain the optimal decision. An experimental result of handwritten numeral recognition is presented, supporting the effectiveness of our method. 1.

We present a practical learning algorithm for a combined classifier with a reject decision, by extending the original generalised probabilistic descent (GPD) algorithm to use a reject option. Our algorithm simultaneously minimizes the classification error and the reject risk under any reject conditions, in order to design an optimum combined classifier with a reject decision, rather than an optimum reject decision function. We demonstrate that our combined classifier is superior to a conventional classifier with a reject option by using error-reject characteristic.

We present an extended version of dynamic link matching for the recognition of all objects similar to a stored model. With a given complex image containing multiple objects, our system can detect all of those matched with a model regardless of their type of geometric transformation and distortion. This is achieved by allowing simultaneous strengthening of multiple bundles of interlayer links (dynamic links) when local features in the model have more than one counterpart in the image layer. To extract the matched regions, in which dynamic links are established as geometrically invariant mappings, we propose a neural system based on edge dynamics in the edge-mapping parameter space.

In software development, it has been pointed out that software engineers must pay attention to software requirements definition. One of the important problems in software engineering is to rationalize the processes from requirements definition to design. Computer tools are most useful and efficient for this purpose. This paper proposes a computer-aided software design system (CASDS), which supports software engineers with a series of structural modeling. As is well-known in systems planning, structural modeling helps to extract concepts from many fuzzy requirements. This system contains three structural modeling methods. They are used 1) to determine functional terms from fuzzy software requirements, 2) to obtain modules by structuring the functions with respect to the data flows, and 3) to make a program skeleton by imposing control flows on the functional elements obtained by breaking down the modules.

By the development of pluviometric radar systems in recent years, spacial distributions of precipitation are available as a time series of digital images. The prediction of the movement and deformation of them is necessary to prevent flooding and to make good use of water resources. A rainfall is greatly influenced by the ground configuration (orographic effect). This paper proposes a time series prediction method in which the movement is estimated by Kalman filter and the ground configuration influence is taken in account as increasing rate and velocity potential. In order to determine the performance of this method, experiments were carried out using practical radar data.