{"ID":5937621,"CreatedAt":"2026-07-07T03:14:33.014478982Z","UpdatedAt":"2026-07-08T07:18:47.304504717Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.04149","arxiv_id":"2607.04149","title":"Beyond Scene Priors: Fine-Grained Traffic Scene Reasoning with Benchmarking and Query-Guided Small-Object Focus","abstract":"In safety-critical traffic scenarios, answering complex questions relies on minute, localized visual cues. However, standard Multimodal Large Language Models (MLLMs) tend to over-attend to backgrounds, overwhelming crucial small objects during visual-language alignment, a failure mode we term 'critical evidence dilution.' Furthermore, existing visual question answering (VQA) datasets rarely expose this flaw, as they lack large-scale, distractor-heavy evaluations that require pinpointing local evidence. To bridge this evaluation and architecture gap, we introduce the Fine-Grained Traffic Reasoning Benchmark (FGTR-Bench) and the Text-Guided Small-Object Reasoning MLLM (TSR-MLLM). FGTR-Bench comprises 40,236 single-image Multiple-Choice Questions (MCQs) created via multi-agent generation, consistency checks, and expert audits, alongside a disjoint 4,947-sample blind test split. To resolve evidence dilution, TSR-MLLM, built on Qwen3-VL-4B, uses a query-conditioned Text-Guided Small-Object Focus (TG-SOF) map. Applied once at the decoder boundary, the map adds sparse Top-K gated residuals to the most question-relevant vision slots while leaving text tokens unchanged. Together with lightweight decoder adaptation, TSR-MLLM preserves single-pass inference without external detectors or image re-encoding. Under matched settings, TSR-MLLM outperforms the strongest 4B baseline by 2.1 points on FGTR-Bench (74.1% overall), with larger gains on evidence-local tracks. Furthermore, it remains competitive on DriveQA-V (CARLA Signs) under greedy decoding without task-specific fine-tuning.","short_abstract":"In safety-critical traffic scenarios, answering complex questions relies on minute, localized visual cues. However, standard Multimodal Large Language Models (MLLMs) tend to over-attend to backgrounds, overwhelming crucial small objects during visual-language alignment, a failure mode we term 'critical evidence dilutio...","url_abs":"https://arxiv.org/abs/2607.04149","url_pdf":"https://arxiv.org/pdf/2607.04149v1","authors":"[\"Waikit Xiu\",\"Qiang Lu\",\"Zian Wang\",\"Xinjie Yang\",\"Zhiwei Chen\",\"Chen Sun\",\"Xiying Li\"]","published":"2026-07-05T07:27:02Z","proceeding":"cs.CV","tasks":"[\"cs.CV\"]","methods":"[\"Large Language Model\",\"Language Model\"]","has_code":false}