Matthew J. Burlick
Assistant Teaching Professor
Department of Computer Science
3141 Chestnum St.
Philadelphia, PA 19104, USA
Office: University Crossings 137
Office hours: By appointment.
BioI am currently an Assistant Teaching Professor in the Computer Science Department at Drexel University in Philadelphia.PA.
TeachingBelow is a list of the courses I am or have taught at Drexel
Generally speaking my research interest include topics in Computer Vision and Machine Learning, in particular understanding video semantics and interpreting human annotation. Below is a list of projects I have worked on, both past and present:Semantics-Based Video Indexing using Tracked Objects
Using object trackers, we discover video semantics for tasks including video segmentation, story discovery, video matching and searching. Some parts of this work are included in my current Thesis work.Multi-Modal Fusion for Voice Activity Detection
Using classification from both acoustic and visual modalities, we provide a novel late-fusion classifier that aides in instantaneous voice activity detection (VAD) in the presence of low acoustic SNR.SCAR Tracking
By tracking on small, salient face patches, we provide a face tracker that is more robust to noise such as illumination and occlusion. Furthermore our algorithms provide mechanisms to detection failure, resulting in object re-acquisition, dynamic model update, or true loss-of-target.Shotgun Learning
A novel semi-supervised learning method using temporally consistent data to label streams instead of individual images to build a training setX-Reality
Real-time visualization of tracking objects in a virual environment. This allows the user to interact with the environment in a previously impossible manner. Doing so we can build behavorial modals for environmentsBeamforming
Using hydrophone data and statistical correction methods, we determine the direction of arrival of ship traffic and establish a guarenteed capture probability.Ground Truths of Subjective Tasks
Acquired ground truth data from several subjects for use in creating a probabilistic model for the segmentation of several videos.