PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge For Detailed Image Descriptions

• URL: http://arxiv.org/abs/2510.19060v1
• Date: Tue, 21 Oct 2025 20:30:20 GMT
• Title: PoSh: Using Scene Graphs To Guide LLMs-as-a-Judge For Detailed Image Descriptions
• Authors: Amith Ananthram, Elias Stengel-Eskin, Lorena A. Bradford, Julia Demarest, Adam Purvis, Keith Krut, Robert Stein, Rina Elster Pantalony, Mohit Bansal, Kathleen McKeown,
• Abstract summary: PoSh is a metric for detailed image description that uses scene graphs as structured rubrics to guide LLMs-as-a-Judge.<n>PoSh is replicable, interpretable and a better proxy for human raters than existing metrics.<n>We show that PoSh achieves stronger correlations with the human judgments in DOCENT than the best open-weight alternatives.
• Score: 55.95282725491425
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: While vision-language models (VLMs) have advanced into detailed image description, evaluation remains a challenge. Standard metrics (e.g. CIDEr, SPICE) were designed for short texts and tuned to recognize errors that are now uncommon, such as object misidentification. In contrast, long texts require sensitivity to attribute and relation attachments and scores that localize errors to particular text spans. In this work, we introduce PoSh, a metric for detailed image description that uses scene graphs as structured rubrics to guide LLMs-as-a-Judge, producing aggregate scores grounded in fine-grained errors (e.g. mistakes in compositional understanding). PoSh is replicable, interpretable and a better proxy for human raters than existing metrics (including GPT4o-as-a-Judge). To validate PoSh, we introduce a challenging new dataset, DOCENT. This novel benchmark contains artwork, paired with expert-written references, and model-generated descriptions, augmented with granular and coarse judgments of their quality from art history students. Thus, DOCENT enables evaluating both detailed image description metrics and detailed image description itself in a challenging new domain. We show that PoSh achieves stronger correlations (+0.05 Spearman $\rho$) with the human judgments in DOCENT than the best open-weight alternatives, is robust to image type (using CapArena, an existing dataset of web imagery) and is a capable reward function, outperforming standard supervised fine-tuning. Then, using PoSh, we characterize the performance of open and closed models in describing the paintings, sketches and statues in DOCENT and find that foundation models struggle to achieve full, error-free coverage of images with rich scene dynamics, establishing a demanding new task to gauge VLM progress. Through both PoSh and DOCENT, we hope to enable advances in important areas such as assistive text generation.

Related papers

• Generating an Image From 1,000 Words: Enhancing Text-to-Image With Structured Captions [33.440447854396446]
  We train the first open-source text-to-image model on long structured captions.<n>To process long captions efficiently, we propose DimFusion.<n>We also introduce the Text-as-a-Bottleneck Reconstruction (TaBR) evaluation protocol.
  arXiv  Detail & Related papers  (2025-11-10T09:25:25Z)
• TokBench: Evaluating Your Visual Tokenizer before Visual Generation [75.38270351179018]
  We analyze text and face reconstruction quality across various scales for different image tokenizers and VAEs.<n>Our results show modern visual tokenizers still struggle to preserve fine-grained features, especially at smaller scales.
  arXiv  Detail & Related papers  (2025-05-23T17:52:16Z)
• Towards Visual Text Grounding of Multimodal Large Language Model [74.22413337117617]
  We introduce TRIG, a novel task with a newly designed instruction dataset for benchmarking text-rich image grounding.<n>Specifically, we propose an OCR-LLM-human interaction pipeline to create 800 manually annotated question-answer pairs as a benchmark.<n>A comprehensive evaluation of various MLLMs on our proposed benchmark exposes substantial limitations in their grounding capability on text-rich images.
  arXiv  Detail & Related papers  (2025-04-07T12:01:59Z)
• A Novel Evaluation Framework for Image2Text Generation [15.10524860121122]
  We propose an evaluation framework rooted in a modern large language model (LLM) capable of image generation. A high similarity score suggests that the image captioning model has accurately generated textual descriptions. A low similarity score indicates discrepancies, revealing potential shortcomings in the model's performance.
  arXiv  Detail & Related papers  (2024-08-03T09:27:57Z)
• BRIDGE: Bridging Gaps in Image Captioning Evaluation with Stronger Visual Cues [47.213906345208315]
  We propose BRIDGE, a new learnable and reference-free image captioning metric. Our proposal achieves state-of-the-art results compared to existing reference-free evaluation scores.
  arXiv  Detail & Related papers  (2024-07-29T18:00:17Z)
• Image Textualization: An Automatic Framework for Creating Accurate and Detailed Image Descriptions [30.08331098481379]
  We propose an innovative framework termed Image Textualization (IT) IT automatically produces high-quality image descriptions by leveraging existing multi-modal large language models (MLLMs) and multiple vision expert models. We show that LLaVA-7B, benefiting from training on IT-curated descriptions, acquire improved capability to generate richer image descriptions.
  arXiv  Detail & Related papers  (2024-06-11T17:37:45Z)
• FINEMATCH: Aspect-based Fine-grained Image and Text Mismatch Detection and Correction [66.98008357232428]
  We propose FineMatch, a new aspect-based fine-grained text and image matching benchmark. FineMatch focuses on text and image mismatch detection and correction. We show that models trained on FineMatch demonstrate enhanced proficiency in detecting fine-grained text and image mismatches.
  arXiv  Detail & Related papers  (2024-04-23T03:42:14Z)
• Do LVLMs Understand Charts? Analyzing and Correcting Factual Errors in Chart Captioning [90.13978453378768]
  We introduce a comprehensive typology of factual errors in generated chart captions. A large-scale human annotation effort provides insight into the error patterns and frequencies in captions crafted by various chart captioning models. Our analysis reveals that even state-of-the-art models, including GPT-4V, frequently produce captions laced with factual inaccuracies.
  arXiv  Detail & Related papers  (2023-12-15T19:16:21Z)
• Visual Data-Type Understanding does not emerge from Scaling Vision-Language Models [31.69213233651326]
  We introduce the novel task of Visual Data-Type Identification. An extensive zero-shot evaluation of 39 vision-language models (VLMs) shows a nuanced performance landscape.
  arXiv  Detail & Related papers  (2023-10-12T17:59:30Z)
• FACTUAL: A Benchmark for Faithful and Consistent Textual Scene Graph Parsing [66.70054075041487]
  Existing scene graphs that convert image captions into scene graphs often suffer from two types of errors. First, the generated scene graphs fail to capture the true semantics of the captions or the corresponding images, resulting in a lack of faithfulness. Second, the generated scene graphs have high inconsistency, with the same semantics represented by different annotations.
  arXiv  Detail & Related papers  (2023-05-27T15:38:31Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.