3D Scenes for Visual Attention Data Collection
Conner Eugene Klawon
Fixation datasets are commonly used for machine learning. By studying how humans actually look at objects we can help teach computers to look at objects similarly to humans. Additionally, understanding the way human attention works allows us to know and predict biases formed in human attention, such as a blind spots, through computer vision. However this requires teaching computers vision skills first. The method of doing this requires fixation datasets from human subjects. Currently most research is done using 2D fixation datasets, which is where this project looks to step in. The future is computer vision in our 3D world, and it therefore needs fixation datasets with 3 dimensions of space. In this project, we build a dataset of 3D scenes which can be use to extract human fixation data.
Van Tam Nguyen
Primary Advisor's Department
Stander Symposium project, College of Arts and Sciences
"3D Scenes for Visual Attention Data Collection" (2022). Stander Symposium Projects. 2472.