The researches for automatic novel composition started on the basis of the successful results in knowledge-engineering/science, and today is one of interesting topics in the knowledge society or the intelligence environment age. The important element is a scenario, which is necessary to determine the novel structure as the systematical scheme with interactive relationships among events in spatio-temporal view. Also, a story is introduced to represent the physical time-sequence among events. The event order in scenario is not the same as that in story. In this paper, we address an effective support function for composing scenarios, which makes beginners to write easily with help of externalization. Two different approaches “top-down-like divergence” and “bottom-up-like convergence” are comparatively supported in our creative thinking process. Our key concepts for scenario composition are “flexible manipulation for creative work” and “externalization of various types of descriptions/specifications under two approaches”.

This paper proposes a map construction system for disturbed disaster areas based on ant colony systems. In our system, evacuees with mobile devices record roads they take while evacuating to a shelter as data collection agents. Then, at the shelter, recorded road information is gathered and restructured into a safe-road map. In this restructuring, information is weighted in accordance with the ant colony system. By this mechanism, we treat the reliability of information on road segments with respect to their safeness for evacuation. In the simulation experiment, we can observe that our system can follow a changing situation and provide an up-to-minute safe-road map to a shelter in a disaster area.

There are a number of issues which are involved with organizing a conference. Among these issues, assigning conference-papers to reviewers is one of the most difficult tasks. Assigning conference-papers to reviewers is automatically the most crucial part. In this paper, we address this issue of paper-to-reviewer assignment, and we propose a method to model the reviewers, based on the matching degree between the reviewers and the papers by combining a preference-based approach and a topic-based approach. We explain the assignment algorithm and show the evaluation results in comparison with the Hungarian algorithm.

Currently it is not difficult for us to find out necessary news articles timely from Web information space. However, it is not always easy to search relevant news articles for a time-dependent sequence of events, related to some interesting topic. In this paper, we propose an aggregation method for collecting news articles, related to a time-dependent sequence of events for a particular topic. Our idea is to represent the topic as a pair of two different nouns and a set of related news articles on a time-dependent change of topic.

In collaborative learning, learners discuss with other learners by exchanging their utterances. In order to represent effective utterances for knowledge acquisition timely, we propose a detection method of focusing utterances which are useful for understanding about their discussion. Focusing utterances are detected according to utterance types and understanding knowledge of learners. Focusing utterances are represented emphatically in the interface so as to distinguish them from normal utterances. Experimental result showed that our method could detect more than 60% of focusing utterances correctly.

So far, Automatic Signature Verification (ASV) approaches using a threshold-based decision have depended on one feature set for distance measure and a threshold on this distance measure for verification. The best performance that can be reached in this case is the one obtained by using the best feature set (bfs). In this paper, we introduce a new decision making approach for ASV that uses Multi-Sets of Features (MSF). The MSF provides higher performance than that obtainable by using the bfs, with better forgery detection. The improvement is seen to be significant because it recovers some lost effectiveness and can add it to that of the bfs. This gain in effectiveness is highly desirable when we deal with signatures of high value documents.

In this paper, we aim to capture dependency among high correlated attributes in a sliding window of data streams at any time. Naive solutions cannot address combinatorial explosion problems. In order to generate highly correlated sets, we propose a candidate-generation rule based on Jaccard distance, and a generation method based on min-hashing functions to estimate the Jaccard distance without requiring to count supports explicitly. For mining frequent ratios, we introduce a concept of ratio range to reduce the candidate ratio sets efficiently. Additionally, the downward closure property of ratio patterns is also available to prune candidate sets. Our extensive experiments on both real and synthetic datasets verify the efficiency of our algorithms.

Recently, power consumption of server computers is one of the most important topics. Although DVS (dynamic voltage scaling) can reduce power consumption, this method is only used at idle time or low load computation time, generally. In this paper, we propose a task scheduling algorithm for reducing power consumption especially at high load computation time. DVS is applied to tasks that are not in critical path, which can reduce more power consumption by introducing a task selection mechanism without increasing the makespan of task graphs. Experimental results show that our algorithm can reduce about 9.1% and 12.8% of power consumption on 4 and 8 processors respectively on average in comparison with the existing method.

Aggregate information is important for data stream processing systems, especially, to get any user-specified time window queries and evolution analysis over data stream. To solve this problem, in this paper, we propose an integrated structure for managing summarized information of snapshots under geometry timeframe, which facilitates analyzing the temporal evolving of data stream. By using our method, the cost of update operations and the errors can be controlled within acceptable scope. Evaluation shows the our structure can respond to arbitrary time window aggregate queries within small errors, efficiently.

This paper proposes an approach to extracting features of origami works which are recognized from facing page images of instructions described as origami drill books. The origami work is presented as a set effaces (polygons). First, in order to associate geometric shapes of objects with knowledge, an origami data model based on design knowledge is proposed. Therefore, the set of faces is classified into clusters which indicate a certain meaning in origami design. The cluster is called "origami molecule ". Next, the relationships among the molecules are defined. The relationships are not only connected relationships but also symmetrical relationships and equivalence relationships. By using these relationships, it is possible to comprehend twin parts of origami work, for example, two forefeet of the four-limbed creature. Finally, a method for extraction of the design knowledge from crease patterns is proposed.

This paper proposes a method for representing overlapping-faces of 3D virtual origami in order to support beholder's recognition of its conformation. Generally, an origami model is constructed by planar polygons corresponding to faces of origami. Therefore, when an origami model is displayed, multiple faces (polygons) on the same plane are probably recognized as one face. Our proposed method moves such overlapping-faces apart slightly by rotating polygons along a rotation axis determined from shapes and relationships of faces. As a consequence, beholders can recognize the conformation of origami correctly. This method can be applied to the user interface of a system we develop to recognize folding operations from origami drill books.

In this paper, we focus on behavioral decisions for empty transport vehicles e.g., they should wait on their place or go to other places. Effective behaviors of empty vehicles enable to decrease waiting times of customers. Our algorithm to acquire such effective behaviors is based on transport experiences of vehicles (i.e., history). It means that vehicles adjust to their city environment by learning trends of transport demands (e.g., amounts and directions). Our algorithm consists of learning and abstraction stages. In the learning stage, vehicles acquire low-level rules of actions at an intersection. In the abstraction stage, vehicles acquire high-level rules of actions in a region. Finally, we report simulation results and show the effectiveness of our algorithm

Route planning is one of the important tasks for transport systems. Appropriate policies for route selections improve not only profitability of transport companies but also convenience of customers. Traditional ways for establishing the policies depend on manual efforts based on statistical data of transports. Moreover, traditional route planning techniques are reactive, i.e., an optimization based on information provided in advance. It is difficult for the manual policies and the reactive planning techniques to adjust dynamic changes of transport trends for customers such as amount and direction of transport demands, i.e., drivers of transport vehicles must follow the policies provided in advance. Therefore, in this paper we show how the proactive route planning based on expected rewards for transport systems can be modeled as a reinforcement learning problem. And, we show how agents as transport vehicles acquire their policies for route selection autonomously and dynamically. The learning ability of transport trends enables transport vehicles to foresee the next destination which provides high rewards. Finally, we report simulation results and make the effectiveness of our proposal strategy clear

In these years, a dynamic transport service which is called demand-bus system by using the contemporary realtime location technologies is focused by public administrations. The most important feature of the services is on-demand processing; i.e., customers can call transport vehicles anytime and get on/off anywhere. However, in the existing circumstances, the demand-bus system is still on the developing stage. In this paper, we classify the demand-bus system into two types: semi-demand and full-demand, on the basis of bus station types. Moreover, we propose methods to solve the real-time problems, (i.e., assignment and scheduling problems), in the demand-bus system by using anticipated future routes of transport vehicles. Finally, we report simulation results and consider effectiveness and possibility of the demand-bus system in an urban area.

Shadows of moving objects often obstruct robust visual tracking. In this paper, we present a car tracker based on a hidden Markov model/Markov random field (HMM/MRF)-based segmentation method that is capable of classifying each small region of an image into three different categories: vehicles, shadows of vehicles, and background from a traffic-monitoring movie. The temporal continuity of the different categories for one small region location is modeled as a single HMM along the time axis, independently of the neighboring regions. In order to incorporate spatial-dependent information among neighboring regions into the tracking process, at the state-estimation stage, the output from the HMMs is regarded as an MRF and the maximum a posteriori criterion is employed in conjunction with the MRF for optimization. At each time step, the state estimation for the image is equivalent to the optimal configuration of the MRF generated through a stochastic relaxation process. Experimental results show that, using this method, foreground (vehicles) and nonforeground regions including the shadows of moving vehicles can be discriminated with high accuracy.

This paper describes a framework to recognizing and recreating folding process of origami based on illustrations of origami drill books. Illustration images acquired from origami books are motley and not sequenced. Moreover, the information obtained from 2D illustrations is so superficial and incomplete that the folding operations cannot be determined uniquely. To solve these problems, a highly flexible and reliable recognition mechanism is proposed. The paper additionally includes the content as follows. Firstly, an algorithm for revising the positions of folding operations extracted from illustrations is proposed so as to make our recognition approach more reliable. Secondly, the outline of the methods which enable feasible folding operations to be generated based only on superficial and incomplete information extracted from illustrations is described. Finally, some updating procedures are proposed to maintain consistency of data (called internal model) which record the transformation of origami models in 3D virtual space during a folding process. Several examples that prove the validness of proposed algorithms/methods are also given in this paper.

In the collaborative learning environment, there are various kinds of information, such as faces of students and learning materials, and the interface which shows these information effectively is required. If the information is provided effectively, students would communicate with others smoothly and progress the learning efficiently. When we study in the group, we tend to focus on what we are interested in or what we thought to be important in order to get meaningful information efficiently. Therefore, our objective is to construct a user-oriented interface which reflects preferences or characteristics of individual students and only provides meaningful information for students. In this paper, the authors propose factors that are needed as the interface of collaborative learning in order to support effective and smooth learning. Then, they introduce the example of user-oriented interface in which students are able to focus on a particular student in the group during the discussion.

Shadows of moving objects often obstruct robust visual tracking. We propose an HMM-based segmentation method which classifies in real time each pixel or region into three categories: shadows, foreground, and background objects. In the case of traffic monitoring movies, the effectiveness of the proposed method has been proven through experimental results.

We address a method for the refinement process of road information based on the generation and verification mechanism of hypotheses to our parallel map recognition, which is organized by a bottom-up approach. First, the hierarchical structure is organized between the generation task and verification task with a view of making the mechanism applicable, and then a generation task in the upper layer controls the corresponding verification task in the lower layer. If available processors should be managed effectively, it is possible to improve the recognition ratio by means of the cooperative interpretation among the related hypotheses. Also, we implemented the prototype system on a parallel computer in order to demonstrate the effectiveness of our approach.

To extract address information automatically from house map images is an important subject for the construction of house databases which support various kinds of regional services for residents. The extraction procedure of home-address numbers can usually be implemented by character recognition techniques. However, it is difficult for many character recognition methods to distinguish distorted or noise-overlapped home-address numbers precisely. To cope with this difficulty, we address a complementation method to infer indeterminate home-address numbers, using already-established home-address numbers of neighboring house blocks. Our method first estimates the address difference among home-address numbers on the basis of a house-block layout pattern, and then infers the indeterminate home-address number from adjacently-surrounding house blocks. We also evaluate our method through experiments.

It is important to provide digitized manuscripts of old literature (in page image form) and their electronic text (in full-text form), with an automatic reference mechanism between the images and the text, on the Internet. As an essential step for creating such an automatic reference system, this paper describes the issue of extracting character areas from page images of old handwritten manuscripts. Page images of old manuscripts are usually terribly dirty and considerable large in size. To overcome the first problem, we propose a new effective method for separating characters from noisy background, since conventional threshold selection techniques are inadequate to cope with the image where the gray levels of the character parts are overlapped by that of the background. To solve the second problem, we propose an approach based on a downscaled image and a recursive labeling method for word extraction. This approach is suitable for large size images because it has the advantage of saving memory and reducing processing time.

In pattern recognition, parallelization, in which an image is partitioned into partial images and each partial image is processed in parallel, is effective for improving processing efficiency. However, the boundary and synchronization problems become drawbacks to the processing efficiency and recognition ratio. For these problems, a recognition strategy based on the information propagation between the same level of neighboring processing data (i.e. horizontal propagation) and between different levels of processing data (i.e. vertical propagation) is successful. We discuss a pipeline-like recognition strategy based on information propagation.

This paper describes an approach for recognition the folding process of origami by interpreting a series of graphics images. Our approach is based on a 3D internal model created during the recognition process. The output of image analysis is matched to a 2D projection of the model such that the contents of folding operations can be identified. The recognized folding operations are used to update the model for further interpretation and recognition. The effectiveness in our approach is illustrated by means of folding process "cicada".

Addresses a parallel recognition paradigm to extract road information from urban map images. The main objective is to make clear the applicability between data parallelization and function parallelization in the road extraction process. We introduce the multilayer partitioned blackboard model. This model manages relationships among the corresponding partitioned segments in different layers of the blackboard so as to perform easily the cooperation among different processing procedures. Namely, individually partitioned segments in each layer are applicable to implement the data parallelization, while the relationships between corresponding partitioned segments in different layers support the cooperative control mechanism of function parallelization effectually.

To interpret motions by analyzing a series of illustrations is an interesting topic. This paper describes an approach for interpreting essential folding operations in illustrated books of origami. Since a folding operation such as "folding up" can be defined by a rotative surface, axis and direction, we proposed the algorithms to identify arrows and dashed lines which represent rotative surfaces, axes and directions in a typical book of origami. The arrow detection process consists of the arrow-head detection and arc detection. The former is accomplished by applying the distance transformation and border following to an input image, while the latter is composed of two phases: all paths which would correspond to arcs are found out in the thinned and approximated image first, an evaluative function is then applied to them to determine arcs. Dashed lines with different patterns are detected using a nearest-neighbor clustering method. A folding operation is finally interpreted based on the identified graphical primitives and the spatial relationships among them. Experimental results on several test images are presented. The effectiveness of our approach in generating meaningful interpretations of origami diagrams is clear from these results.

Document knowledge plays a very important role in many document image understanding methods. In these methods, the document knowledge is utilized to classify/extract individual item data interpretatively from paper-based sheets as a kind of document model: these days, this knowledge is specified into the document image understanding system in advance. In this paper, we propose an experimental method to acquire the layout knowledge automatically from sample document images. In particular, we focus on the acquisition of the layout of business cards. Our idea is to generate the layout knowledge of business cards from a predefined logical structure, which is used as a kind of meta-knowledge to interpretatively generate the layout knowledge of given business cards.

Map recognition is one of the most interesting methods to distinguish meaningful information automatically from map images and to construct resource data in information systems such as GISs (geographic information systems). However, it is difficult to identify individual pieces of information because the map components mutually intersect or are overlayed, and because the properties of the map components are not well defined. In this paper, we propose a method of extracting character strings from color urban map images. The characteristic of our method is to validate the identified character strings with map composition rules. Our recognition process is composed of two phases: extraction of character strings and classification of character strings. The extraction phase is organized by a bottom-up approach, based on measurement/estimation among pixels, while the classification phase is composed by means of a top-down approach, based on interpretation/validation among map components.

Until today, techniques for modeling and managing the time varying characteristics of entities in database have been proposed. These techniques are mainly classified into two categories: time stamp method and version method. These methods limit the representation/manipulation capabilities because the change of class membership for an object is not well specified and the transitional states must be related mutually along the system defined time axis. We propose a method which is basically derived from the object oriented paradigm, for modeling time varying characteristics of entities so as to be free from the above two limitations in the traditional methods. We introduce two representation media: core and surface. The core represents invariant features of entity, while the surface denotes variant features of entity. A snapshot of an entity is represented as an aspect, which is a database instance composed of the core and surfaces.

We address the knowledge representation for understanding table-form documents. In particular, we focus on the knowledge for establishing the correctness of recognized results in addition to the traditional syntactic/constructive-oriented understanding framework. In order to attain to this subject, we introduce the semantic information to specify domains and relationships among individual items definitely, which is complementary to the syntactic/constructive information available in the traditional framework.

The object-oriented model is very applicable to represent various phenomena in the real world from an entity-interaction point of view. The successful modeling methods have been developed on many current topics. However, it is not sufficient to model a group of interrelated objects and dynamic actions of individual objects effectively under this paradigm. We introduce the concept of field in addition to the notion of object with respect to constructing the cooperative environment for objects, and then discuss our modeling method based on the object and field. Our modeling method is successful to construct a group of objects through the field, and to organize a hierarchical structure among objects by looking upon the field as an abstract object conceptually. Also, we explain a property adaptation mechanism of objects in fields with respect to the group activity of objects.>

Many approaches have reported that knowledge-based layout recognition methods are very successful in classifying the meaningful data from document images automatically. However, these approaches are applicable to only the same kind of documents because they are based on the paradigm that specifies the structure definition information in advance so as to be able to analyze a particular class of documents intelligently. In this paper, the authors propose a method to recognize the layout structures of multi-kinds of table-form document images. For this purpose, the authors introduce a classification tree to manage the relationships among different classes of layout structures. The authors' recognition system has two modes: layout knowledge acquisition and layout structure recognition. In the layout knowledge acquisition mode, table-form document images are distinguished according to this. Classification tree and then the structure description trees which specify the logical structures of table-form documents are generated automatically. While, in the layout structure recognition mode, individual item fields in the table-form document images are extracted and classified successfully by searching the classification tree and interpreting the structure description tree.>

We propose a framework to manage interrelated tasks concurrently and successively, and address the issues of event handling and task scheduling: the event handler organizes interrelated tasks from a static point of view, and the task scheduler controls them on the basis of the dynamic behavior properties. In order to design our flexible control mechanism, we introduce the state transition graph and AND-OR graph in the event handler, whereas in the task scheduler the timestamp ordering method is applied to. Additionally, we describe the failure handling, concerning the cooperation between the event handler and task scheduler.>

The authors address visual interface and virtual manipulation for full text management. Their objective is to investigate the retrieval/reference mechanism of computer-readable full texts as if paper-form books were being manipulated directly in the real world. This mechanism is derived basically from the object-linking facility of hypersystems. Of course, the facility must be applicable to the manipulation of a great deal of electronic-formed books, like the traditional information retrieval systems, through hypersystems are suitable to only private filing applications.>

A framework of four-layer recognition processes is proposed for understanding documents, and a knowledge representation method adaptable to the understanding of table-form documents is addressed. Although Y. Nakano et al. (1986) looked upon the recognition of multi-kinds of table-form documents as an important subject from a practical point of view, they could not report any successful approach because their knowledge was based only on the physical coordinate data. In the approach presented, this recognition issue was solved, using both the classification tree based on the physical characteristics and the structure description tree based on the logical characteristics. At least, it is not so difficult to classify various kinds of documents into appropriate document classes since table-form documents are well designed on the basis of vertical and horizontal line segments. However, it is not easy in the case of the other documents because the geometric and spatial characteristics of documents are not well specified. It is necessary to investigate the application techniques for the other documents from the viewpoint of the knowledge representation.>

The main objective of document understanding is to extract and classify the meaningful data automatically from documents. Some researches, concerning this issue, have already been reported. However, these methods are not always successful because the recognition procedures analyze document images on the basis of only physical coordinate values of compositive items. This paper proposes a more advanced method based on the spatial relationships among neighboring segments of compositive items, in addition to the geometric aspects. The knowledge about documents is not a single layer, but organized as multi-level layers: knowledge about layout structures, knowledge about item sequences and knowledge about item properties. Three kinds of knowledge are not only specified hierarchically, but also interrelated mutually between the layout recognition, item recognition and character recognition procedures.>

The authors propose a fundamental framework for the editing facility applicable to uniform manipulation for various kinds of data. The basic principle is to separate the object scheme and the data instance from the editing data extractively. This separation method makes it possible to specify various kinds of data under explicitly structured forms and then manipulate them with the uniform access paths. The editing model has an object-oriented feature based on the relational data model. The design concept of the editing mechanism is described in detail. Integration and functionality issues are addressed.>