John MacCormick's research

My main area of research is computer vision—that is, getting computers to understand images and video—but I enjoy working on a wide variety of problems in computer science. For example, I've also worked on large-scale distributed data storage systems, and pair programming techniques for undergraduate computer science education. Here are a few of the questions my research has attempted to address over the years:

• How can we track a person's finger accurately enough with a video camera to allow fluent human-computer interaction without using any type of touchscreen, mouse, or pointing device? (This is work from my PhD thesis, later published as chapter 7 of the book Stochastic Algorithms for Visual Tracking, and in the conference paper Partitioned sampling, articulated objects, and interface-quality hand tracking.)
• When undergraduate students work in pairs on programming assignments in an introductory computer science course, how should pairs be assigned? Randomly? Or is it better to pair strong students with other strong students and less able students with other less able students? Or do pairs of mixed ability lead to better outcomes? (This question is tackled in a SIGCSE conference paper, The Benefits of Pairing by Ability.)
• Suppose you are trying to store, in a data center, millions of files on behalf of thousands of customers. To guard against data loss, each file should have, say, three copies stored on different computers in the data center. How do you decide where to store the copies of each file, in such a way that the file can be efficiently retrieved later, while keeping the load on each computer balanced? (A possible solution is provided by an article in the ACM Transactions on Storage, Kinesis: A New Approach to Replica Placement in Distributed Storage Systems.)

To get a more complete picture of my research, please take a look at my publications and patents. You may also be interested in the list of talks I have given, and a separate page describing research with undergraduate students.

My most recent project, called MultiCam, examines the technical and user interface challenges in providing support for multiple cameras in video chat software such as Skype. For more details, and to download the MultiCam software, please check out the MultiCam project page.