CustomSight: Enhancing LLM-Powered Visual Assistance for Blind Individuals using Goal-Directed Dynamic Filters

  • Adil Rahman
  • , Rifat Rahman Khan
  • , Jonggi Hong
  • , Stephanie Valencia
  • , Seongkook Heo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

LLM-powered assistive technologies (ATs) have enabled blind and visually impaired (BVI) users to query personalized, goal-oriented information about their visual environment. However, the accuracy of system responses depends heavily on well-framed, query-relevant images, which can be difficult for BVI users to capture. We present CustomSight, an LLM-powered AT that helps BVI users effectively query visual information by providing task-aware, real-time guidance to frame the camera and automatically capture images when relevant content is in view. When a user issues a query, CustomSight generates a Dynamic Filter-a custom pipeline that encodes logic tied to the user’s intent, monitors the live feed, and triggers context-aware feedback and image capture. The captured image is sent to the LLM to fetch accurate visual information.

Original languageEnglish
Title of host publicationUIST Adjunct 2025 - Proceedings of the 38th Annual ACM Symposium on User Interface Software and Technology
EditorsAndrea Bianchi, Elena Glassman, Shengdong Zhao, Jeeeun Kim, Ian Oakley, Wendy E. Mackay
ISBN (Electronic)9798400720369
DOIs
StatePublished - 27 Sep 2025
Event38th Annual ACM Symposium on User Interface Software and Technology, UIST 2025 - Busan, Korea, Republic of
Duration: 28 Sep 20251 Oct 2025

Publication series

NameUIST Adjunct 2025 - Proceedings of the 38th Annual ACM Symposium on User Interface Software and Technology

Conference

Conference38th Annual ACM Symposium on User Interface Software and Technology, UIST 2025
Country/TerritoryKorea, Republic of
CityBusan
Period28/09/251/10/25

Keywords

  • Accessibility
  • Assistive Technology
  • Blind and Low Vision
  • LLMs

Fingerprint

Dive into the research topics of 'CustomSight: Enhancing LLM-Powered Visual Assistance for Blind Individuals using Goal-Directed Dynamic Filters'. Together they form a unique fingerprint.

Cite this