iTrace Infrastructure - Enabling Eye Tracking in Developer Environments

Software developers work with many source code files but each of these files typically is comprised of thousands of lines of code. They constantly flip through multiple files while reading and debugging code. Current state-of-the-art eye tracking software works well on fixed stimuli such as images but does not support tracking of developer gazes on these semantically and syntactically rich source code artifacts. Since source code files are really large, we cannot feasibly draw AOIs around each element to map gaze data to those regions. iTrace, our eye tracking infrastructure solves these problems making it easier for one to get started on conducting eye tracking studies in software engineering and related fields.

Onion UML Visualization

Using OnionUML notation to create focus + context views in UML.

Visually Correcting Gaze Fixations

A tool that allows visually correcting groups of fixations for short stimuli.

Towards Automatically Correct Fixations in Source Code

A gradient descent algorithm that automatically corrects fixations. Evaluation is done on a set of manually corrected fixations.