Cooperative Distributed Image Understanding Systems

协作分布式图像理解系统

• 批准号: 08408010
• 负责人: MATSUYAMA Takashi
• 金额: $ 21.25万
• 依托单位: KYOTO UNIVERSITY
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (A)
• 财政年份: 1996
• 资助国家: 日本
• 起止时间: 1996 至 1998
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-08408010/
• 关键词: Image Understanding; Cooperative Distributed processing; Cooparative Distributed Vision; Multi-Agent System; Behavior Recognition; Object Tracking; Real Time Video Processing; Selective Attention Mechanism; 分散協調型画像理解; 分散観測; Appearance Sphere; 協

In this study we consider the cooperative distributed processing (multi-agent system) as a system architecture to realize robust and flexible image understanding systems, and demonstrate its effectiveness by developing practical cooperative distributed image understanding systems. The results of the study can be summarized as follow :1. The types of cooperative distributed processing for image understanding are classified into the following categories : (a) Distributed Sensing Systems (b) Distributed Analysis Systems ((b-1) Functional Distribution (b-2) Algorithm Distribution (b-3) Cooperative Spatial Reasoning).2. A multi-agent region segmentation system, where each image region is regarded as an agent3. A fixed viewpoint pan-tilt camera to acquire ominidirectional images and a camera calibration software to model an off-the-shelf active camera as a fixed viewpoint pan-tilt-zoom camera4. A real time object detection and tracking system using a fixed viewpoint pan-tilt-zoom camera5. A cooperative object tracking system with multiple distributed fixed viewpoint pan-tilt-zoom cameras6. A cooperative distributed behavior recognition from multiple video data7. A wide area object tracking system consisting a group of distributed observation stations without common visual fields

本研究将协作分布式处理（多智能体系统）作为实现鲁棒、灵活图像理解系统的系统架构，并通过开发实际的协作分布式图像理解系统来验证其有效性。研究结果可以总结为：1。用于图像理解的协作分布式处理的类型分为以下几类：(a)分布式传感系统(b)分布式分析系统（（b-1）功能分布（b-2）算法分布（b-3）协作空间推理）。一种多agent区域分割系统，其中每个图像区域被视为一个agent。一种用于获取全方位图像的固定视点泛倾斜相机和一种将现成的活动相机建模为固定视点泛倾斜变焦相机的相机校准软件。一种采用固定视点平移变焦相机的实时目标检测与跟踪系统。多分布式固定视点平移-倾斜变焦相机协同目标跟踪系统基于多视频数据的协同分布式行为识别方法[j]。由一组分布式观测站组成的广域目标跟踪系统

项目成果

T.Wada: "Appearance Sphere : Background Model for Pan-Tilt-Zoom Camera" Proc.of 13th International Conference on Pattern Recognition. A-718-A-722 (1996)

T.Wada：“外观球体：平移倾斜变焦相机的背景模型”第 13 届国际模式识别会议论文集。

和田俊和,浮田宗伯,松山隆司: "視点固定型パン・チルト・ズームカメラとその応用" 電子情報通信学会論文誌. Vol.J81-D II No.6. 1182-1193 (1998)

Toshikazu Wada、Sohaku Ukita、Takashi Matsuyama：“固定视点平移/倾斜/变焦相机及其应用”，电子、信息和通信工程师学会卷 J81-D II 第 6 期（1998 年）。

T.Matsuyama: "Cooperative Spatial Reasoning for Image Understanding" International Journal of Pattern Recognition and Artificail Intelligence. 11, No.1. 205-227 (1997)

T.Matsuyama：“图像理解的协作空间推理”国际模式识别和人工智能杂志。

T.Wada: "Fixed Viewpoint Pan-Tilt-200m Camera and its Applications (in Japanese)" Trans.of IEICE. J81-D-II,No.6. 1182-1193 (1998)

T.Wada：“固定视点 Pan-Tilt-200m 摄像机及其应用（日语）”Trans.of IEICE。

T.Wada and T.Matsuyama: "Appearance Sphere : Background Model for Pan-Tilt-Zoom Camera" Proc.of 13th International Conf.on Pattern Recognition. A718-A722 (1996)

T.Wada 和 T.Matsuyama：“外观球体：平移倾斜变焦相机的背景模型”第 13 届国际模式识别会议论文集。