Enregistré dans:
Détails bibliographiques
Auteurs principaux: Ramachandran, Akhil, Arun, Ankit, Shenoy, Ashish, Harpale, Abhay, Jayakumar, Srihari, Chatterjee, Debojeet, Moslehpour, Mohsen, Chuang, Pierce, Lu, Yichao, Bhardwaj, Vikas, Heidari, Peyman
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:https://arxiv.org/abs/2602.13479
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912904297578496
author Ramachandran, Akhil
Arun, Ankit
Shenoy, Ashish
Harpale, Abhay
Jayakumar, Srihari
Chatterjee, Debojeet
Moslehpour, Mohsen
Chuang, Pierce
Lu, Yichao
Bhardwaj, Vikas
Heidari, Peyman
author_facet Ramachandran, Akhil
Arun, Ankit
Shenoy, Ashish
Harpale, Abhay
Jayakumar, Srihari
Chatterjee, Debojeet
Moslehpour, Mohsen
Chuang, Pierce
Lu, Yichao
Bhardwaj, Vikas
Heidari, Peyman
contents Video Large Language Models (Video LLMs) have shown remarkable progress in understanding and reasoning about visual content, particularly in tasks involving text recognition and text-based visual question answering (Text VQA). However, deploying Text VQA on wearable devices faces a fundamental tension: text recognition requires high-resolution video, but streaming high-quality video drains battery and causes thermal throttling. Moreover, existing models struggle to maintain coherent temporal context when processing text across multiple frames in real-time streams. We observe that text recognition and visual reasoning have asymmetric resolution requirements - OCR needs fine detail while scene understanding tolerates coarse features. We exploit this asymmetry with a hybrid architecture that performs selective high-resolution OCR on-device while streaming low-resolution video for visual context. On a benchmark of text-based VQA samples across five task categories, our system achieves 72% accuracy at 0.49x the power consumption of full-resolution streaming, enabling sustained VQA sessions on resource-constrained wearables without sacrificing text understanding quality.
format Preprint
id arxiv_https___arxiv_org_abs_2602_13479
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle GLIMPSE : Real-Time Text Recognition and Contextual Understanding for VQA in Wearables
Ramachandran, Akhil
Arun, Ankit
Shenoy, Ashish
Harpale, Abhay
Jayakumar, Srihari
Chatterjee, Debojeet
Moslehpour, Mohsen
Chuang, Pierce
Lu, Yichao
Bhardwaj, Vikas
Heidari, Peyman
Computer Vision and Pattern Recognition
Human-Computer Interaction
Video Large Language Models (Video LLMs) have shown remarkable progress in understanding and reasoning about visual content, particularly in tasks involving text recognition and text-based visual question answering (Text VQA). However, deploying Text VQA on wearable devices faces a fundamental tension: text recognition requires high-resolution video, but streaming high-quality video drains battery and causes thermal throttling. Moreover, existing models struggle to maintain coherent temporal context when processing text across multiple frames in real-time streams. We observe that text recognition and visual reasoning have asymmetric resolution requirements - OCR needs fine detail while scene understanding tolerates coarse features. We exploit this asymmetry with a hybrid architecture that performs selective high-resolution OCR on-device while streaming low-resolution video for visual context. On a benchmark of text-based VQA samples across five task categories, our system achieves 72% accuracy at 0.49x the power consumption of full-resolution streaming, enabling sustained VQA sessions on resource-constrained wearables without sacrificing text understanding quality.
title GLIMPSE : Real-Time Text Recognition and Contextual Understanding for VQA in Wearables
topic Computer Vision and Pattern Recognition
Human-Computer Interaction
url https://arxiv.org/abs/2602.13479