Research

Feature spaces for recognition and categorization

I am interested in understanding how image features are extracted and used for behavioral tasks. My projects involve modeling shape and texture spaces used in object categorization and face recognition, and using texture-synthesis models to investigate the features used in scene perception tasks.

• Ehinger, K. A. & Altschuler, E. L. (2011). What did the early American presidents really look like? Gilbert Stuart portraits as a "Rosetta Stone" to the pre-photography era. Perception.
  project page
• Ehinger, K. A. & Oliva, A. (2008). Characterizing the shape and texture of natural objects using Active Appearance Models. Poster presented at the Vision Sciences Society annual meeting, May 11, 2008.

Scene typicality and categorization

What makes for a “typical” beach or a “typical” street scene? Are some views of scenes more typical than others? I am interested in modeling the features which distinguish typical views and exemplars to better understand scene categorization.

• Ehinger, K. A., Haggerty, K. M., & Oliva, A. (2010). Canonical views of scenes depend on the shape of the space. Poster presented at Object Perception, Attention, & Memory, Nov, 18, 2010.
• Xiao, J., Hays, J., Ehinger, K. A., Oliva, A., & Torralba, A. (2010). SUN Database: Large scale scene recognition from abbey to zoo. In Proc. 23rd IEEE Conference on Computer Vision and Pattern Recognition, 3485 - 3492.
  project page

Modeling visual search in natural scenes

We characterize the image features that are thought to drive visual search in scenes -- saliency, features of the search target, and the global context of the scene -- and examine the ability of models based on these features to predict human eye movements over a large collection of scenes.

• Ehinger, K. A.*, Hidalgo-Sotelo, B.*, Torralba, A., & Oliva, A. (2009). Modeling search for people in 900 Scenes: A combined source model of eye guidance. Visual Cognition, 17, 945-978.
  project page
• Ehinger, K. A.*, Hidalgo-Sotelo, B.*, Torralba, A., & Oliva, A. (2009). Modeling search for people in 900 Scenes: A combined source model of eye guidance. Poster presented at the MIT Scene Understanding Symposium, Jan 30, 2009.