A Comparative Analysis of Recent Advances in Deep Learning for Image Captioning
Pertanika Proceedings · 2025
DOI: 10.47836/pp.1.5.005 ↗Kurz gefasst
Fachartikel in der Zeitschrift Pertanika Proceedings, 2025. Verfasst von Bhargavi Polepalli und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Image captioning uses visual perception combined with natural language processing to increase accessibility and support uses including automated monitoring. Transformer models are among the sophisticated methods produced by conventional approaches grounded on recurrent neural networks (RNNs) and convolutional neural networks (CNNs). These models increase the contextual accuracy and grammatical consistency of output captions since they are known to control long-range dependencies. Emphasizing transformer-based developments, this article evaluates and compares modern deep learning methods for image captioning via a comprehensive analysis. To improve caption production, the work investigates models combining Swin Transformers, diffusion models, and depth fusion. Comparative measures, such as BLEU, ROUGE, METEOR, and CIDEr, are used to demonstrate the performance benefits made by these approaches over conventional models, with a focus on efficiency. This paper demonstrates the considerable influence of recent deep learning algorithms on image quality and utility.
Zitieren
APA
Polepalli, B., Sekharamantry, P. K., & Rao, K. S. (2025). A Comparative Analysis of Recent Advances in Deep Learning for Image Captioning. Pertanika Proceedings. https://doi.org/10.47836/pp.1.5.005
Vancouver
Polepalli B, Sekharamantry PK, Rao KS. A Comparative Analysis of Recent Advances in Deep Learning for Image Captioning. Pertanika Proceedings. 2025. doi:10.47836/pp.1.5.005
BibTeX
@article{Polepalli2025,
author = {Polepalli, Bhargavi and Sekharamantry, Praveen Kumar and Rao, Konda Srinivasa},
title = {A Comparative Analysis of Recent Advances in Deep Learning for Image Captioning},
journal = {Pertanika Proceedings},
year = {2025},
doi = {10.47836/pp.1.5.005}
}
EUVIMED