Deep Learning Approach for Image Captioning
Vrushali R. Pagire, Ashish Kale, Murthy S. Chavali · 2025
Automatic image captioning is the process of generating textual descriptions or captions for images based on their content. It represents a machine learning challenge that integrates computer vision (for understanding image content) and natural language processing (for generating text). This field is relatively new and rapidly growing, with ongoing research introducing innovative techniques to improve results. However, significant progress is still needed to achieve outcomes that match human-level performance. In this paper, the proposed method LSTM-GAN is implemented for image captioning. The dataset used for the work is MS COCO. The performance of the proposed method is evaluated by calculating the BLEU (Bilingual Evaluation Understudy) score. The BLEU score value is between 0.6 to 0.7 which shows good results.