Clinic Projects

Please click on a link below to view the Harvey Mudd College Computer Science Clinic projects for the corresponding time period.

Clinic Projects for 2007-2008

A Web-Based PDF Management and Organization Solution

Client
Bluebeam Software, Inc.

Faculty Advisor
Professor Christine Alvarado

Student Team
Jay Markello (Project Manager), Eduardo Ruvalcaba, Joe Simmons, Scott Triglia
The Clinic team researched and developed a system to organize and manage large sets of PDFs (Portable Document File documents). The proposed system enables users to find the document they need quickly and efficiently by searching through data within the file and user supplied data about the file. The team measured the system's success by how much it simplifies the process of finding a particular file.

A Web-Based Implementation of a Reservation and Scheduling System for Efficient Per-Seat Pricing for Air Taxi Service

Client
D4 Networks, LLC

Faculty Advisor
Professor Zachary Dodds

Student Team
Andrew La Motte-Mitchell (Project Manager), Chris Alvino, Corey Hebert, Steven Sloss
The goal of this project was to develop a web application highlighting the flexibility and convenience of charter air travel. This application enables passengers and charter operators to pool their resources in order to keep travel competitive with commercial airlines. The team's efforts enable an optimization algorithm from a 2006-2007 HMC Math Clinic to act as an engine for a charter-booking and delivery business.

Visualizing Proof Search

Client
Fair Isaac Corporation

Faculty Advisor
Professor Robert Keller

Student Team
Mike Buchanan (Project Manager-Fall), Michael Ernst, Phil Miller (Project Manager-Spring), Chris Roberts
Fair Isaac deals with large knowledge bases in a variety of their lines of business. They are developing an automated theorem prover in the natural deduction framework to build on these data sets. The scale of the proofs and their attendant search spaces make textual proof display and analysis of the prover's operation unreasonable. Thus, the Clinic team has developed a visualization system which greatly eases development efforts. It provides a structured display of the theorem prover's search space and a programmable command-line interface which gives the developer significantly more flexibility than a conventional debugger would allow.

CPU and GPU Based Image Processing for Digital Photographers

Client
Microsoft Corporation

Faculty Advisor
Professor Elizabeth Sweedyk

Student Team
Lucy Abramyan (Project Manager), Morgan Conbere, Ellen Kephart, Lilia Markham, Matt McKnett
Through over a century of research and practice, professional film photographers developed visual aesthetics for compelling images, with "looks" pleasing for the human eye. These techniques are often at odds with the non-perceptual, analytical image processing algorithms applied to digital photographs today. This project examines how these worlds can be united by developing aesthetically-inspired algorithms and prototypes (on the CPU and GPU) for perceptual saturation enhancement, soft focus simulation, and locally-modified high-dynamic range image processing.

FLO Analysis Tool (FLOAT)

Client
QUALCOMM, Inc.

Faculty Advisor
Professor Geoff Kuenning

Student Team
T. Andrew Glass (Project Manager), Andrew Pienkos, Michael Roberts, Kris Karr (Fall)
MediaFLO is a mobile television technology that has been developed by QUALCOMM. Since FLO (Forward Link Only) signals behave unpredictably when subjected to real-world conditions, it can be very challenging for engineers to diagnose issues with MediaFLO transmissions. Our project expedited this process by developing a series of applications that will aid in the collection and analysis of MediaFLO signal data while enabling communication between engineers in the field and in a central location.

Development and Characterization of a Real-Time Video Streaming System

Client
RealNetworks, Inc.

Faculty Advisor
Professor Christopher Stone

Student Team
Thomas Barr (Project Manager), Sameer Sontakey, Peter Mawhorter, Daniel Rozeboom
The team developed a live video delivery system that utilizes the upload capacity of clients in a peer-to-peer (p2p) system. While p2p systems have become popular for on-demand content delivery in recent yers, no successful live streaming system has ever been deployed commercially. The team developed a novel system, and will fully test its behavior in a simulated network.