CISE Research Instrumentation: A Laboratory for Interactive Applications for Computational Vision and Language

9818322 Dickinson，SvenStevenson，SuzanneNew Brunswick，NJ 08854- 8010计算视觉和视觉交互应用实验室该研究仪器使以下研究项目成为可能：- 视觉对象识别及其在残疾人辅助机器人中的应用，-基于模型的估计，跟踪和识别以观察人类活动，-基于自然语言生成的评估交互系统，以及-解决交互式指令中的歧义。该奖项支持建立，在罗格斯大学计算机科学和工程的持续研究，并有助于购买美能达3-D激光扫描仪/对象建模，游牧技术XR 4000机械臂和立体视觉升级，视觉一个视频跟踪平台，四个STW Linux的PC的，以及语言数据联盟（文本语料库的来源）的成员。 该实验室将由四位首席研究员领导，并会由四个涵盖电脑视觉及自然语言处理范畴的合作研究项目共用：（1）将物体识别及语言理解功能整合于一个供残疾人士使用的机器人系统;（2）在基于模型的框架中结合物体和事件的联合收割机估计、跟踪和识别的综合视觉系统的开发;（3）用于评估自动生成的指令的视频分析系统的设计;以及（4）自然语言指令的自动歧义消解过程的开发。 在每种情况下，所产生的技术都嵌入到人机界面中，从而朝着实现更自然的交互系统的目标迈出了一步。

9818322Dickinson, SvenStevenson, SuzanneNew Brunswick, NJ 08854-8010A Laboratory for Interactive Applications of Computational Vision and LanguageThis research instrumentation enables research projects in:- Visual Object Recognition and its Applications to Robotic Aids for the Disabled,- Model-Based Estimation, Tracking, and Recognition for Observing Human Activities,- Evaluating Interactive Systems Based on Natural Language Generation, and- Resolving Ambiguities in Interactive Instructions.This award supports established, on-going research in computer science and engineering at Rutgers University and contributes to the purchase of a Minolta 3-D laser scanner/object modeler, a Nomadics Technologies XR4000 Robotic Arm and Stereo Vision Upgrade, a Vision One Video Tracking Platform, four STW Linux-based PC's, and a membership in the Linguistic Data Consortium (a source of text corpora). The equipment forms the basis for a new Laboratory for Interactive Applications of Computational Vision and Language, directed by the four PIs, and will be shared among four collaborative research projects spanning the areas of computer vision and natural language processing: (1) the integration of object recognition and language understanding in a robotic system for the disabled; (2) the development of integrated vision systems that combine estimation, tracking, and recognition of both objects and events in a model-based framework; (3) the design of a video analysis system for evaluating automatically generated instructions; and (4) the development of automatic ambiguity resolution procedures for natural language instructions. In each case, the resulting technologies are embedded within human-computer interfaces, thereby taking a step toward achieving the goal of more natural interactive systems.