Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{174181,
        author = {V ADI LAKSHMI SRAVYA VUTUKURI and MUPPARAJU KEERTHI and KOYI PRADEEP CHOUDHARY and KURRI GOVINDA REDDY and BURRA LALITHA RAJESWARI},
        title = {Visual Question Answering and Image Caption Generation Using Deep Learning},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {11},
        number = {10},
        pages = {3493-3501},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=174181},
        abstract = {The study develops two computer vision applications through AI technology which include Visual Question Answering (VQA) and Image Caption Generation. A VQA system operates by combining computational vision methods with natural lan- guage processing capabilities to both read pictures and produce appropriate user responses. The system makes use of ResNet-
50 to extract image features and BERT to process textual questions thus providing precise context-based answers. The data augmentation process benefits from the similar image search API and users gain a better experience through JavaScript-based image enhancement.
With deep learning methods the Image Caption Generation system produces meaningful text descriptions that describe im- ages. The system integrates VGG16 for visual feature extraction method together with LSTM-based sequence modeling to produce natural language captions. The system receives education through extensive databases containing image-caption pairs that allow it to create coherent captions that align with the visual context. The system usability receives enhancement through additional features which include similar image retrieval and image en- hancement techniques.
The training methods include Adam optimization and Cosine Annealing Learning Rate Scheduler and Early Stopping to achieve both accuracy and efficient learning in both models. Real-world performance assessments indicate that VQA achieved 90.11},
        keywords = {Visual Question Answering, Image Caption Generation, ResNet-50, BERT, VGG16, LSTM,},
        month = {March},
        }