Bhanuja Ainary | Searcharxiv

arXiv2026

Audo-Sight: AI-driven Ambient Perception Across Edge-Cloud for Blind and Low Vision Users

Despite advances in assistive technologies, Blind and Low-Vision (BLV) individuals continue to face challenges in understanding their surroundings. Delivering concise, useful, and timely scene descriptions for ambient perception remains a long-standing accessibility problem. To address this, we introduce Audo-Sight, an AI-driven assistive system across Edge-Cloud that enables BLV individuals to perceive their surroundings through voice-based conversational interaction. Audo-Sight employs a set of expert and generic AI agents, each supported by dedicated processing pipelines distributed across edge and cloud. It analyzes user queries by considering urgency and contextual information to infer the user intent and dynamically route each query, along with a scene frame, to the most suitable pipeline. In cases where users require fast responses, the system simultaneously leverages edge and cloud processing pipelines. The edge generates an initial response quickly, while the cloud provides more detailed and accurate information. To overcome the challenge of seamlessly combining these outputs, we introduce the Response Fusion Engine, which fuses the fast edge response with the more accurate cloud output, ensuring timely and high-accuracy response for the BLV users. Systematic evaluation shows that Audo-Sight delivers speech output around 80% faster for urgent tasks and generates complete responses approximately 50% faster across all tasks compared to a commercial cloud-based solution -- highlighting the effectiveness of our system across edge-cloud. Human evaluation of Audo-Sight shows that it is the preferred choice over GPT-5 for 62% of BLV participants with another 23% stating both perform comparably.

cs.DC↗

arXiv2025

Audo-Sight: Enabling Ambient Interaction For Blind And Visually Impaired Individuals

Visually impaired people face significant challenges when attempting to interact with and understand complex environments, and traditional assistive technologies often struggle to quickly provide necessary contextual understanding and interactive intelligence. This thesis presents Audo-Sight, a state-of-the-art assistive system that seamlessly integrates Multimodal Large Language Models (MLLMs) to provide expedient, context-aware interactions for Blind and Visually Impaired (BVI) individuals. The system operates in two different modalities: personalized interaction through user identification and public access in common spaces like museums and shopping malls. In tailored environments, the system adjusts its output to conform to the preferences of individual users, thus enhancing accessibility through a user-aware form of interaction. In shared environments, Audo-Sight employs a shared architecture that adapts to its current user with no manual reconfiguration required. To facilitate appropriate interactions with the LLM, the public Audo-Sight solution includes an Age-Range Determiner and Safe Query Filter. Additionally, the system ensures that responses are respectful to BVI users through NeMo Guardrails. By utilizing multimodal reasoning, BVI-cognizant response editing, and safeguarding features, this work represents a major leap in AI-driven accessibility technology capable of increasing autonomy, safety, and interaction for people with visual impairments in social settings. Finally, we present the integration of Audo-Sight and SmartSight, which enables enhanced situational awareness for BVI individuals. This integration takes advantage of the real-time visual analysis of SmartSight, combined with the extensive reasoning and interactive capabilities of Audo-Sight, and goes beyond object identification to provide context-driven, voice-controlled assistance in dynamic environments.

cs.HC↗