Ray Rimey

Projects

This page provides some more detail about many of my research projects. Click on a picture for its full-size version. A blue diamond denotes papers that are more important to me, . All my papers are accessible from my Publications page, where they are listed by year.


Activity Recognition, LM, 2004-present.

Several projects developed methods to model, learn, detect and recognize activities.

Related Publications:


Sketch Representations of Activity, LM, 2006-2008.

This project explored how sketchs and diagrams can be used to represent complex activity and how analysts might use sketch-based tools to reason about activity.

Related Publications:


Zoomable User Interfaces (ZUIs) and Lens Tools, LM, 1997-2004.

Several projects explored how zoomable user interface (ZUI) and lens concepts can benefit future analysis tools. One project examined analysis tasks involving a large amount and variety of visual work materials, techniques for externalizing analysis tasks on a ZUI canvas, and the benefits of the resulting new analysis environment. These projects built on Professor Ben Bederson's Piccolo, Jazz and Pad++ toolkits for zoomable graphics. We developed lenses in numerous tools: Piccolo, Jazz, Pad++, OpenMap, JLoox, WebELT, Swing.

Related Publications:


Intelligent Mobile Sensor System (IMSS), LM, 1995-1997

The IMSS project developed a fully-autonomous indoor mobile robot system that drives around and inspects 55-gallon drums stored in warehouses. The project culminated with two month-long field tests inside warehouses containing drums of low-level radioactive and hazardous waste at the DOE's Fernald, OH and Idaho Falls, ID locations.

Related Publications:


Unmanned Ground Vehicle / Demo II, LM, 1993-1997

The UGV / Demo II program, begun in 1992, developed and matured those navigation and automatic target recognition technologies critical for the development of supervised, autonomous ground vehicles capable of performing military scout missions with a minimum of human over-sight. The program culminated with a highly successful series of field exercises performed by soldiers at Ft. Hood, Texas over three weeks in May-June 1996.

Related Publications:


Control of Selective Perception (PhD thesis), Department of Computer Science, University of Rochester, 1990-1993

My PhD thesis developed methods for selecting a sequence of camera viewpoints and selecting algorithms to process the acquired images in order to answer a question about a scene. My advisor was Professor Chris Brown.

Related Publications:


Applications of Hidden Markov Models to Computer Vision, Department of Computer Science, University of Rochester, 1988-1989

This project examined the use of hidden Markov models for some computer vision problems, including using them to drive the motion sequences of a camera.

Related Publications:


Applications of Neural Networks to Computer Vision, Nestor Inc., 1985-1988

I developed the performance evaluation subsystem of the Nestor Development System product (NDS-1000), a developer toolkit for solving pattern classification problems. I also created demo applications of the NDS to three computer vision applications. Nestor existed from 1983 to 2009, and developed products for character recognition, fraud detection, and red-light camera systems.

Related Publications:


Recognizing Models in Range Data (MS thesis), Department of Electrical Engineering, University of Rhode Island, 1983-1985

I developed a maximum likelihood approach to segmenting 3D range data and matching 3D object models. My advisor was Professor Fernand Cohen, who is now at Drexel University. I also built the low-level software to enable acquisition and manipulation of raw range data from a sensor built by another student.

Related Publications:


Extracting Wireframe Building Models from Aerial Images, Department of Electrical and Computer Engineering, Carnegie Mellon University, 1982-1983

This was an undergraduate project during my senior year and the summer before. I developed a system that hypothesized matches of a flat-roofed N-sided building model to a line/junction segmentation of an aerial image of an urban area, and used a straight-line Hough transform to find additional predicted, but weaker, building edges in the original image data. My project sponsor was Professor Takeo Kanade.

Related Publications: