Research

Sidewalk at BC

Art and Vision
Prof. Stella Yu

We investigate human perception through a series of eye tracking experiments which reveal what visual information is essential for a human viewer to perform a detection or recognition task. Ongoing projects are:

  • scene categorization
  • change detection
  • motion perception

The experimental data provide insights into what features and how they are computed in certain visual routines.

Understanding both the peculiarity of human vision and the universality of computational constraints allows us to better computation in the areas of:

  • image and video compression
  • brightness and color perception
  • scene layout inference
  • motion depiction

Data mining
Prof. Sergio Alvarez 

Data mining is data analysis plus algorithmics. The objective is to develop methods to efficiently extract useful information from large data sets. I have made contributions to the analytical foundations of this field, as well as to applications within medicine and personalized information systems. Please note that I take great care to safeguard any private personal information that is needed in my work. My work in data mining with colleagues and students at WPI has focused on the paradigm of association rules. I am also collaborating with medical colleagues from the U. of Massachusetts Medical School. My most recent work involves the discovery of statistically significant patterns in data arising in human sleep studies and in surgical oncology.


EagleEyes Project
Profs. James Gips, Peter Olivieri, William Ames

EagleEyes is a technology that allows the user to control the computer by moving his eyes or head. The computer senses changes in the angles of the eyes through electrodes placed around the eyes. Currently dozens of children and young adults who are non-speaking and have complex disabilities are using EagleEyes on a regular basis at the Boston College Campus School and at other facilities in the U.S. and England. This project involves work in human-computer interaction, psychophysiology, assistive technology, circuit design, applications of computers and the internet in education.

 

Ground Truth Dataset and Benchmarks for Mid-Level Vision
Prof. David Martin

Machine vision systems can now do amazing things: Reading irises and faces, helping to drive autonomous cars in real environments, locating and measuring anatomical structures in medical scans -- these are just a few examples of capabilities that have emerged in recent years. Special-purpose domains still mark the limit of our success, however. The goal of human-level machine vision is still out of reach because the solutions found to these problems do not require the machine to understand the rich structure of visual information.

It is essential to take an empirical approach to the problem of visual perception. The primary goal of this project is to build a dataset of ground truth image annotations that provides the perception of scenes at the level of surfaces, objects, and basic 3D scene geometry. This dataset will be unprecedentedly rich and detailed, providing precisely the information and representations needed to bring general purpose capabilities to machine vision systems. A secondary goal of this project is to create the associated benchmarks and methodologies for evaluating machine systems with respect to the ground truth data.