@inproceedings{guo-etal-2025-learning-describe,
    title = "Learning to Describe Implicit Changes: Noise-robust Pre-training for Image Difference Captioning",
    author = "Guo, Zixin  and
      Sun, Jiayang  and
      Wang, Tzu-Jui Julius  and
      Radman, Abduljalil  and
      Pehlivan, Selen  and
      Cao, Min  and
      Laaksonen, Jorma",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/name-variant-enfa-fane/2025.findings-emnlp.537/",
    doi = "10.18653/v1/2025.findings-emnlp.537",
    pages = "10125--10145",
    ISBN = "979-8-89176-335-7",
    abstract = "Image Difference Captioning (IDC) methods have advanced in highlighting subtle differences between similar images, but their performance is often constrained by limited training data. Using Large Multimodal Models (LMMs) to describe changes in image pairs mitigates data limits but adds noise. These change descriptions are often coarse summaries, obscuring fine details and hindering noise detection. In this work, we improve IDC with a noise-robust approach at both data and model levels. We use LMMs with structured prompts to generate fine-grained change descriptions during data curation. We propose a Noise-Aware Modeling and Captioning (NAMC) model with three modules: Noise Identification and Masking (NIM) to reduce noisy correspondences, Masked Image Reconstruction (MIR) to correct over-masking errors, and Fine-grained Description Generation (FDG) to produce coherent change descriptions. Experiments on four IDC benchmarks show that NAMC, pre-trained on our large-scale data, outperforms streamlined architectures and achieves competitive performance with LLM-finetuned methods, offering better inference efficiency."
}