Takao Miyamoto

Recognition and Textual Description of Human Activities by Mobile Robot

In this paper, we propose a method for recognizing human actions and objects and translating them into natural language text. First, 3D environmental map is constructed by accumulating range maps captured from a 3D range sensor mounted on a mobile robot. Then, pose of a person in the scene is estimated by fitting articulated cylindrical model and also object is recognized by matching 3D models. On condition that the person handles some objects, interaction with the object is classified. Finally, using conceptual model representing human actions and related objects, a natural language expression which is most suitable to explain the persons action is generated.

Development and case study of trend analysis software based on FACT-Graph

This article proposes text mining software to analyze FACT-Graph, and describes a case study using the software. FACT-Graph is a trend graph which visualizes what kinds of topic exist and shows the changes in trends in time-series text data. However, FACT-Graph itself does not have enough environments to analyze trends although it provides clues for a trend. In order to resolve this problem, we developed the software called Loopo. This software provides the functions of adding the considerations of the analyst as the keywords, and operating FACT-Graph itself such as moving, adding, and clearing nodes. The system also allows analysts to refer to an information source, keyword information, and network information in order to analyze and consider FACT-Graph. In a case study about criminal trends using the titles of newspaper articles between 1987 and 2007, we confirmed the usability of this software.

Measurement Evaluation of Keyword Extraction Based on Topic Coverage

This paper proposes a method to measure the performance of keyword extraction based on topic coverage. The answer set of a keyword is required to evaluate keyword extraction by methods such as TF-IDF. However, creating an answer set for a large document is expensive. Thus, this paper proposes a new measurement called topic coverage on the basis of the assumption that the keywords extracted by a superior method can express the topic information efficiently. The experiment using the proceedings of a conference shows the feasibility of our proposed method.

Generating Natural Language Annotation from Video Sequences Taken by Handy Camera

In this paper, we propose a method for generating natural language annotation automatically from video sequences taken by handy camera. In general, video image and natural language are totally different form of information. The main challenge is how to bridge the gap between them and to translate one to another. In our method, an appropriate predicate is selected for expressing video contents by combining semantic features extracted from video sequences. It is, however, difficult to select most appropriate words for expressing the contents. Therefore we construct a concept hierarchy of motions and situations to select verbs from coarse to fine. We perform some experiments to test the effectiveness of the proposed method, and confirmed appropriate annotations are generated.

Development of Intelligent Surveillance System Focused on Comprehensive Flow

Surveillance cameras are today a common sight in public spaces and thoroughfares, where they are used to prevent crime and monitor traffic. However, human operators have limited attention spans and may miss anomalies. Here, we develop an intelligent surveillance system on the basis of spatio-temporal information in comprehensive flow of human traffic. The comprehensive flow is extracted from optical flows, and anomalies are identified on the basis of the spatiotemporal distribution. Because our system extracts only a few anomalies from many surveillance cameras, operators will not miss the important scenes. In experiment, we confirmed effectiveness of our intelligent surveillance system.

A Far Road Sign Recognition by Applying Super-Resolution to Extracted Regions from Successive Frames

In the field of ITS technology, many methods have been proposed for detecting and recognizing road signs from car-mounted cameras. Most of these techniques are, however, for only still images. Therefore, the recognition results are affected by occasional changes of light conditions or occlusions. Moreover, though recognition of far placed signs are meaningful for safety driving, these signs are taken in low resolution images which that they are hardly recognition. In this article, we propose a novel technique for recognizing road signs. In this technique, a series of low resolution images of a road sign is extracted by using CONDENSATION from successive frames. Then, a pseudo-high resolution image of the road sign is generated by super-resolution using the series of images. The proposed techniques lead to early recognition results for far road signs even on images which are failed by traditional techniques.

Supporting System for Network Server Monitoring

Networks have complex systems comprised of subsystem components like servers for various kinds of service and DCE (Data Communication Equipment) such as routers and switches. As each system operates constantly, it is hard to know exactly how the whole network operates. The network monitoring system, therefore, is required, in addition to monitoring and checking the network under steady operation, to detect system failures, weigh the situation and investigate into the cause. In this paper, we propose a method that we convert the three kinds of data, data by polling the server externally, data from computer resource of the server, and data from log information of the network server, into an integrated format, and then assemble them as generalized log format, so that we can extract the abnormal events without providing the monitoring system with any data such as key words, in advance. Firstly, we identify the patterns of frequency of the words by text-mining of the log information. Secondly, by means of signal processing of the frequency, periodically appeared patterns are figured out. Lastly, these patterns are associated with each other to presume the cause of the abnormal events. Our proposed method should support network administrators greatly in detecting system failures and investigating into the cause, and also reduce a lot of workload in monitoring log information.