{"slug": "key-frame-reasoning-with-sam3-third-place-solution-for-the-mevis-text-track-of", "title": "Key-Frame Reasoning with SAM3: Third Place Solution for the MeViS-Text Track of the 8th LSVOS Challenge", "summary": "A two-stage, training-free solution using Gemini-3.1 Pro and SAM3 ranked third in the MeViS-Text track of the 8th LSVOS Challenge, achieving J&F, J, F, N-acc., T-acc., and Final scores of 0.761, 0.7367, 0.7852, 0.8333, 0.9755, and 0.856593, respectively. The method decomposes video-level events into instance-level targets, generates key-frame descriptions, and propagates masks bidirectionally using SAM3 on a single NVIDIA GeForce RTX 4090 without task-specific training.", "body_md": "arXiv:2608.17279v1 Announce Type: new\nAbstract: This report presents a two-stage, training-free solution for the MeViS-Text track of the 8th LSVOS Challenge. The task requires a model to localize and segment the object specified by a natural-language expression throughout a video. Such expressions often depend on temporal cues, including actions, interactions, directions, and relative positions. Our first stage uses Gemini-3.1 Pro via API to decompose a video-level event into instance-level targets, select a key frame for each target, and generate a discriminative description aligned with that frame. In the second stage, SAM3-agent produces a pixel-level seed mask on the selected frame, and the SAM3 video tracker propagates the mask bidirectionally through the video. Valid instances are grounded and propagated independently before their frame-wise masks are merged. All local SAM3 processing runs on a single NVIDIA GeForce RTX 4090 without task-specific training or model ensembling. Our method ranked third on the challenge test set, obtaining J&F, J, F, N-acc., T-acc., and Final scores of 0.761, 0.7367, 0.7852, 0.8333, 0.9755, and 0.856593, respectively.", "url": "https://wpnews.pro/news/key-frame-reasoning-with-sam3-third-place-solution-for-the-mevis-text-track-of", "canonical_source": "https://arxiv.org/abs/2608.17279", "published_at": "2026-08-19 04:00:00+00:00", "updated_at": "2026-08-19 04:12:58.575098+00:00", "lang": "en", "topics": ["computer-vision", "artificial-intelligence"], "entities": ["Gemini-3.1 Pro", "SAM3", "MeViS-Text", "8th LSVOS Challenge", "NVIDIA GeForce RTX 4090"], "alternates": {"html": "https://wpnews.pro/news/key-frame-reasoning-with-sam3-third-place-solution-for-the-mevis-text-track-of", "markdown": "https://wpnews.pro/news/key-frame-reasoning-with-sam3-third-place-solution-for-the-mevis-text-track-of.md", "text": "https://wpnews.pro/news/key-frame-reasoning-with-sam3-third-place-solution-for-the-mevis-text-track-of.txt", "jsonld": "https://wpnews.pro/news/key-frame-reasoning-with-sam3-third-place-solution-for-the-mevis-text-track-of.jsonld"}}