Multi-Modal LLM based Image Captioning in ICT: Bridging the Gap Between General and Industry Domain

• URL: http://arxiv.org/abs/2601.09298v1
• Date: Wed, 14 Jan 2026 09:01:46 GMT
• Title: Multi-Modal LLM based Image Captioning in ICT: Bridging the Gap Between General and Industry Domain
• Authors: Lianying Chao, Haoran Cai, Xubin Li, Kai Zhang, Sijie Wu, Rui Xu,
• Abstract summary: This paper proposes a multi-stage progressive training strategy to train a Domain-specific Image Captioning Model (DICModel) in ICT.<n> Experimental results indicate that our DICModel with only 7B parameters performs better than other state-of-the-art models with 32B parameters.
• Score: 10.823938734002288
• License: http://creativecommons.org/licenses/by-sa/4.0/
• Abstract: In the information and communications technology (ICT) industry, training a domain-specific large language model (LLM) or constructing a retrieval-augmented generation system requires a substantial amount of high-value domain knowledge. However, the knowledge is not only hidden in the textual modality but also in the image modality. Traditional methods can parse text from domain documents but dont have image captioning ability. Multi-modal LLM (MLLM) can understand images, but they do not have sufficient domain knowledge. To address the above issues, this paper proposes a multi-stage progressive training strategy to train a Domain-specific Image Captioning Model (DICModel) in ICT, and constructs a standard evaluation system to validate the performance of DICModel. Specifically, this work first synthesizes about 7K image-text pairs by combining the Mermaid tool and LLMs, which are used for the first-stage supervised-fine-tuning (SFT) of DICModel. Then, ICT-domain experts manually annotate about 2K image-text pairs for the second-stage SFT of DICModel. Finally, experts and LLMs jointly synthesize about 1.5K visual question answering data for the instruction-based SFT. Experimental results indicate that our DICModel with only 7B parameters performs better than other state-of-the-art models with 32B parameters. Compared to the SOTA models with 7B and 32B parameters, our DICModel increases the BLEU metric by approximately 56.8% and 20.8%, respectively. On the objective questions constructed by ICT domain experts, our DICModel outperforms Qwen2.5-VL 32B by 1% in terms of accuracy rate. In summary, this work can efficiently and accurately extract the logical text from images, which is expected to promote the development of multimodal models in the ICT domain.

Related papers

• OmniScience: A Large-scale Multi-modal Dataset for Scientific Image Understanding [13.03315906747549]
  We introduce OmniScience, a high-fidelity multi-modal dataset spanning more than 10 major scientific disciplines.<n>We develop a dynamic model-routing re-captioning pipeline that generates dense, self-contained descriptions.<n> pipeline is reinforced with rigorous quality filtering and alignment with human expert judgments.
  arXiv  Detail & Related papers  (2026-02-14T13:08:13Z)
• Text-Guided Semantic Image Encoder [25.15773515839525]
  We propose the Text-Guided Semantic Image (TIE), which generates image representations conditioned on the input text query.<n>TIE-based vision-language models (VLMs) attain superior performance while utilizing only half as many image tiles (tokens)<n>TIE consistently attends to query-relevant regions, enhancing both interpretability and query-specific grounding.
  arXiv  Detail & Related papers  (2025-11-25T19:04:04Z)
• Exploring a Unified Vision-Centric Contrastive Alternatives on Multi-Modal Web Documents [99.62178668680578]
  We propose Vision-Centric Contrastive Learning (VC2L), a unified framework that models text, images, and their combinations using a single vision transformer.<n> VC2L operates entirely in pixel space by rendering all inputs, whether textual, visual, or combined, as images.<n>To capture complex cross-modal relationships in web documents, VC2L employs a snippet-level contrastive learning objective that aligns consecutive multimodal segments.
  arXiv  Detail & Related papers  (2025-10-21T14:59:29Z)
• Text-Visual Semantic Constrained AI-Generated Image Quality Assessment [47.575342788480505]
  We propose a unified framework to enhance the comprehensive evaluation of both text-image consistency and perceptual distortion in AI-generated images.<n>Our approach integrates key capabilities from multiple models and tackles the aforementioned challenges by introducing two core modules.<n>Tests conducted on multiple benchmark datasets demonstrate that SC-AGIQA outperforms existing state-of-the-art methods.
  arXiv  Detail & Related papers  (2025-07-14T16:21:05Z)
• Multimodal Integrated Knowledge Transfer to Large Language Models through Preference Optimization with Biomedical Applications [7.751808693373747]
  MINT (Multimodal Integrated kNowledge Transfer) is a framework that aligns unimodal large decoder models with domain-specific decision patterns from multimodal biomedical data.<n> MINT leverages an upstream multimodal machine learning (MML) model trained on high-quality multimodal data to transfer domain-specific insights to downstream text-only or image-only models.
  arXiv  Detail & Related papers  (2025-05-09T02:28:41Z)
• Leopard: A Vision Language Model For Text-Rich Multi-Image Tasks [62.758680527838436]
  We propose Leopard, an MLLM tailored for handling vision-language tasks involving multiple text-rich images.<n>First, we curated about one million high-quality multimodal instruction-tuning data, tailored to text-rich, multi-image scenarios.<n>Second, we proposed an adaptive high-resolution multi-image encoding module to dynamically optimize the allocation of visual sequence length.
  arXiv  Detail & Related papers  (2024-10-02T16:55:01Z)
• ARMADA: Attribute-Based Multimodal Data Augmentation [93.05614922383822]
  Attribute-based Multimodal Data Augmentation (ARMADA) is a novel multimodal data augmentation method via knowledge-guided manipulation of visual attributes. ARMADA is a novel multimodal data generation framework that: (i) extracts knowledge-grounded attributes from symbolic KBs for semantically consistent yet distinctive image-text pair generation. This also highlights the need to leverage external knowledge proxies for enhanced interpretability and real-world grounding.
  arXiv  Detail & Related papers  (2024-08-19T15:27:25Z)
• VEGA: Learning Interleaved Image-Text Comprehension in Vision-Language Large Models [76.94378391979228]
  We introduce a new, more demanding task known as Interleaved Image-Text (IITC) This task challenges models to discern and disregard superfluous elements in both images and text to accurately answer questions. In support of this task, we further craft a new VEGA dataset, tailored for the IITC task on scientific content, and devised a subtask, Image-Text Association (ITA)
  arXiv  Detail & Related papers  (2024-06-14T17:59:40Z)
• ANCHOR: LLM-driven News Subject Conditioning for Text-to-Image Synthesis [6.066100464517522]
  We introduce the Abstractive News Captions with High-level cOntext Representation dataset, containing 70K+ samples sourced from 5 different news media organizations. Our proposed method Subject-Aware Finetuning (SAFE), selects and enhances the representation of key subjects in synthesized images by leveraging LLM-generated subject weights. It also adapts to the domain distribution of news images and captions through custom Domain Fine-tuning, outperforming current T2I baselines on ANCHOR.
  arXiv  Detail & Related papers  (2024-04-15T21:19:10Z)
• Planting a SEED of Vision in Large Language Model [73.17530130368053]
  We present SEED, an elaborate image tokenizer that empowers Large Language Models (LLMs) with the ability to SEE and Draw at the same time. This version of SEED was trained in 5.7 days using only 64 V100 GPUs and 5M publicly available image-text pairs.
  arXiv  Detail & Related papers  (2023-07-16T13:41:39Z)
• Multimodal Masked Autoencoders Learn Transferable Representations [127.35955819874063]
  We propose a simple and scalable network architecture, the Multimodal Masked Autoencoder (M3AE) M3AE learns a unified encoder for both vision and language data via masked token prediction. We provide an empirical study of M3AE trained on a large-scale image-text dataset, and find that M3AE is able to learn generalizable representations that transfer well to downstream tasks.
  arXiv  Detail & Related papers  (2022-05-27T19:09:42Z)

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.