Capturing Moments: An AI-Powered Image Captioning System

Scene comprehension is a method that automatically generates natural language captions for images that explain the material observed in the images. It integrates knowledge of computer vision with natural language processing. This task has numerous applications in fields such as robotics, human–computer interaction, and multimedia content analysis. Most images have little captions, but we can still understand them in large part without them. So, we chose Flickr 8 K as our system support. The dataset is used to train multiple images and multiple captions. It helps to generate the captions for the raw images. The generated captions aim to accurately describe the image’s content, providing valuable information for various applications such as image retrieval, scene understanding, and accessibility for the visually impaired. In this paper, we applied transfer and deep learning methods such as VGG16, LSTM to propose captions for the images. We used BLEU score as a metric to evaluate our model performance. Our model is performing better when we compared the score of our model with other deep learning models.