Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage

Saehyung Lee, Seunghyun Yoon, Trung Bui, Jing Shi, Sungroh Yoon
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:33815-33832, 2025.

Abstract

Multimodal large language models (MLLMs) excel at generating highly detailed captions but often produce hallucinations. Our analysis reveals that existing hallucination detection methods struggle with detailed captions. We attribute this to the increasing reliance of MLLMs on their generated text, rather than the input image, as the sequence length grows. To address this issue, we propose a multiagent approach that leverages LLM-MLLM collaboration to correct given captions. Additionally, we introduce an evaluation framework and a benchmark dataset to facilitate the systematic analysis of detailed captions. Our experiments demonstrate that the proposed evaluation method aligns better with human judgments of factuality than existing metrics. Moreover, we show that current approaches for enhancing MLLM factuality often fail in hyper-detailed image captioning tasks. In contrast, our approach significantly enhances the factual accuracy of captions, even improving those generated by GPT-4V. Finally, we highlight a limitation of VQA-centric benchmarking by demonstrating that an MLLM’s performance on VQA benchmarks may not correlate with its ability to generate detailed image captions.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-lee25aj, title = {Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage}, author = {Lee, Saehyung and Yoon, Seunghyun and Bui, Trung and Shi, Jing and Yoon, Sungroh}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {33815--33832}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/lee25aj/lee25aj.pdf}, url = {https://proceedings.mlr.press/v267/lee25aj.html}, abstract = {Multimodal large language models (MLLMs) excel at generating highly detailed captions but often produce hallucinations. Our analysis reveals that existing hallucination detection methods struggle with detailed captions. We attribute this to the increasing reliance of MLLMs on their generated text, rather than the input image, as the sequence length grows. To address this issue, we propose a multiagent approach that leverages LLM-MLLM collaboration to correct given captions. Additionally, we introduce an evaluation framework and a benchmark dataset to facilitate the systematic analysis of detailed captions. Our experiments demonstrate that the proposed evaluation method aligns better with human judgments of factuality than existing metrics. Moreover, we show that current approaches for enhancing MLLM factuality often fail in hyper-detailed image captioning tasks. In contrast, our approach significantly enhances the factual accuracy of captions, even improving those generated by GPT-4V. Finally, we highlight a limitation of VQA-centric benchmarking by demonstrating that an MLLM’s performance on VQA benchmarks may not correlate with its ability to generate detailed image captions.} }
Endnote
%0 Conference Paper %T Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage %A Saehyung Lee %A Seunghyun Yoon %A Trung Bui %A Jing Shi %A Sungroh Yoon %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-lee25aj %I PMLR %P 33815--33832 %U https://proceedings.mlr.press/v267/lee25aj.html %V 267 %X Multimodal large language models (MLLMs) excel at generating highly detailed captions but often produce hallucinations. Our analysis reveals that existing hallucination detection methods struggle with detailed captions. We attribute this to the increasing reliance of MLLMs on their generated text, rather than the input image, as the sequence length grows. To address this issue, we propose a multiagent approach that leverages LLM-MLLM collaboration to correct given captions. Additionally, we introduce an evaluation framework and a benchmark dataset to facilitate the systematic analysis of detailed captions. Our experiments demonstrate that the proposed evaluation method aligns better with human judgments of factuality than existing metrics. Moreover, we show that current approaches for enhancing MLLM factuality often fail in hyper-detailed image captioning tasks. In contrast, our approach significantly enhances the factual accuracy of captions, even improving those generated by GPT-4V. Finally, we highlight a limitation of VQA-centric benchmarking by demonstrating that an MLLM’s performance on VQA benchmarks may not correlate with its ability to generate detailed image captions.
APA
Lee, S., Yoon, S., Bui, T., Shi, J. & Yoon, S.. (2025). Toward Robust Hyper-Detailed Image Captioning: A Multiagent Approach and Dual Evaluation Metrics for Factuality and Coverage. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:33815-33832 Available from https://proceedings.mlr.press/v267/lee25aj.html.

Related Material