Theses Doctoral

Enabling Agency in Access to Visual Experiences for Blind Users

Jain, Gaurav

Access to visual information is fundamental to how people navigate the world and make sense of their surroundings. For blind and low vision (BLV) users, however, assistive technologies often reduce complex visual experiences to basic, task-oriented outputs. These systems may help complete specific tasks, but they rarely support agency---the ability to explore, make self-directed decisions, and understand the consequences of one’s actions. This dissertation argues that enabling agency should be a central goal of accessibility design, particularly in contexts involving rich visual environments.

To frame this work, I draw on Doug Church’s model of agency, synthesizing three key conditions: (C1) perceivable options, or knowing what possibilities exist; (C2) meaningful choice, or being able to act in ways that reflect one’s intent; and (C3) clear feedback, or understanding the outcome of one’s actions. I examine how these conditions can be supported across two domains of visual experience: navigation and video watching.

This dissertation presents four projects that address different aspects of these conditions. The first is a qualitative study of BLV users’ needs during open-ended navigation, identifying two types of spatial information (shape and layout) that support C1, and surfacing barriers to C2. The second, StreetNav, repurposes public street cameras to deliver real-time, scene-aware feedback during outdoor navigation, addressing C1 and C3. The third, SceneScout, uses street view imagery and large language models to help users preview unfamiliar environments, supporting C1 and C2. The fourth, Front Row, applies computer vision and spatialized audio to tennis broadcasts, enabling users to access raw gameplay information (C1) and form their own interpretations (C2).

Together, these projects reframe visual accessibility not just as a matter of delivering information, but as a question of enabling BLV users to act, decide, and engage with visual experiences on their own terms.

Files

  • thumbnail for gsas-dissertations-000144.pdf gsas-dissertations-000144.pdf application/pdf 2.61 MB Download File

More About This Work

Academic Units
Computer Science
Thesis Advisors
Smith, Brian Anthony
Degree
Ph.D., Columbia University
Published Here
May 27, 2026

Notes

Computer vision, Assistive technologies, Multimodal interaction, Navigation, Video accessibility