R. Mitchell Parry - Assistant Professor - Computer Science

Mitch Parry, Associate Professor of
Computer Science at Appalachian State University

postal mail:	ASU Box 32133 Boone, NC 28608
phone:	+1 (828) 262-2149
fax:	+1 (828) 262-7000
email:	[email protected]
office:	312K Anne Belk Hall

Fall 2024 Office Hours
MW	3:00 pm -	4:00 pm
TR	2:00 pm -	4:00 pm

Home Schedule CV Publications Google Scholar Student Projects
CS 3435: Data Collection & Visualization, CS 4465/5465: Computer Graphics

Student Research Projects

I welcome the opportunity to work with students on research projects. Below you will find some of the projects I am currently interested in. If any of these interest you or you have another idea you think I may be interested in, please contact me.

Title	Description
Commentator Cancelation	Noise cancelation headphones and even cell phones can actively reduce background noise by listening to the background and subtracting it before playback or transmission. In this project, you will apply the same ideas to remove or reduce the "noise" created by television sports commentators. The goal of the project is to reduce the loudness of the commentator while increasing the loudness of ambient game sounds such as the crowd, referees, coaches, and players. Using the same methodology, you could also remove vocals from a music recording for karaoke performances.
Ranking Evaluation	Sports fans love to debate which team or player would win in a head-to-head competition. Evidence provided includes the results of other head-to-head results and the relative merits of each team. For over 70 years, experts have published their combined opinion in sports opinion polls. More recently, computer algorithms rate teams objectively. In this project, you will evaluate the performance of human polls and computer algorithms at predicting the results of head-to-head competitions. Using the same methodology, you could use the same methodology to evaluate human and computer methods for ranking the social hierarchy of mammals that exhibit agonistic head-to-head interactions, such as baboons.
Musical Snippet Similarity	Currently, applications of music similarity apply at the song, artist, and genre level. For example, Pandora creates playlists of similar songs, allmusic.com recommends similar artists, and a music store organizes music by genre. In this project, you will help describe what it means for short snippets of music to be similar and create a method or methods for determining an objective similarity score that indicates the level of similarity between two snippets.
Music Database Player	Typically, music player applications open one file at a time and provide functionality to play, pause, stop, and seek within that file. Playing a new file requires a short pause in the output of the player while the new file loads into memory. In this project, you will extend open-source software to allow multiple open files and fast (imperceptible) switching between songs.
Win Percentage	With four minutes left in the game your team is ahead by 10 points. As a spectator, is it safe to go home knowing that your team will win? In this project, you will analyze historical game data to determine the frequency of a team winning given the current state of the game. As the lead increases or time decreases the leading team is more likely to win but what are the changes of a come-from-behind victory (or come-from-ahead loss)?
Pause Live Radio	Build an music player that buffers sound from the radio (FM or internet) and allows the user to play, pause, and rewind live audio much like a DVR does for television. After building a buffer, you can skip commercials or other content you don't like.
Digital Audio Program Recorder	Build on the DVR concept for audio. Record live FM or internet audio streams and organize it so that users can skip undesired content (like commercials). Implement algorithms that recognize when the same content is replayed, and cut the content into different songs or radio segments. Automatically create "abridged" versions of the radio content that excludes replays.