{"ID":6023385,"CreatedAt":"2026-07-08T01:00:23.257252134Z","UpdatedAt":"2026-07-10T05:14:55.794445831Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.05859","arxiv_id":"2607.05859","title":"AVA-VLM: Adaptive Visual Attention-Vision Language Model for In-the-Wild Construction Site Monitoring","abstract":"Vision-Language Models (VLMs) are promising for construction-site monitoring, and recent construction-tailored VLMs have primarily adapted pretrained VLMs through direct QA-style fine-tuning from a single global image. We argue that this direct paradigm remains limited for in-the-wild deployment in terms of operational range, reliability under reduced-resolution inputs, and inference efficiency. To address these challenges, we propose AVA-VLM, an Adaptive Visual Attention-Vision Language Model that follows a human-inspired coarse-to-fine reasoning strategy. AVA-VLM first reasons over a low-resolution global image and selectively requests a high-resolution local crop only when detailed inspection is needed, similar to how a human inspector zooms in on hard-to-see yet important areas. We further introduce a region-aware Chain-of-Thought dataset that teaches the model when to inspect, where to crop, and how to use local evidence. Experiments show that AVA-VLM improves reliability under long-distance and reduced-resolution conditions while substantially reducing visual-token usage.","short_abstract":"Vision-Language Models (VLMs) are promising for construction-site monitoring, and recent construction-tailored VLMs have primarily adapted pretrained VLMs through direct QA-style fine-tuning from a single global image. We argue that this direct paradigm remains limited for in-the-wild deployment in terms of operational...","url_abs":"https://arxiv.org/abs/2607.05859","url_pdf":"https://arxiv.org/pdf/2607.05859v1","authors":"[\"Younggun Kim\",\"Taeheon Kim\",\"Youngseo Kim\",\"Seunghee Park\"]","published":"2026-07-07T05:30:13Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Language Model\"]","has_code":false}