Next-generation image captioning: A survey of methodologies and emerging challenges from transformers to Multimodal Large Language Models
Bibliographic record
Abstract
The widespread availability of visual data on the Internet has fueled a significant interest in image-to-text captioning systems. Automated image captioning remains a challenging multimodal analytics task, integrating advances in both Computer Vision (CV) and Natural Language Processing (NLP) to understand image content and generate semantically meaningful textual descriptions. Modern deep learning-based approaches have supplanted traditional approaches in image captioning, leading to more efficient and sophisticated models. The development of attention mechanisms and transformer-based architectures has further enhanced the modeling of both language and visual data. Despite these gains, challenges such as long-tailed object recognition, bias in training data, and shortcomings in evaluation metrics constrain the capabilities of current models. Furthermore, an important breakthrough has been made with the recent emergence of Multimodal Large Language Models (MLLMs). By incorporating textual and visual data, MLLMs provide improved captioning flexibility, generative capabilities, and reasoning. However, these models introduce new challenges, including faithfulness, grounding, and computational cost. Although relatively few studies have comprehensively surveyed these developments, this paper provides a thorough analysis of Transformer-based captioning approaches, investigates the shift to MLLMs, and discusses associated challenges and opportunities. We also present a performance comparison of the latest models on the MS-COCO benchmark and conclude with perspectives on potential future research directions.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.003 | 0.006 |
| Open science | 0.003 | 0.002 |
| Research integrity | 0.002 | 0.003 |
| Insufficient payload (model declined to judge) | 0.008 | 0.004 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".