This repository provides code and resources to build an image caption generation model, which generates natural language descriptions of images. The model uses a pre-trained VGG16 network for feature extraction and an LSTM network for sequence generation.
Overview
Model Architecture
Dataset
The model utilizes two main components:
VGG16 CNN Model for image feature extraction.
LSTM RNN Model for generating captions from the extracted image features.
The VGG16 model extracts a feature vector from an image. The LSTM model then takes this feature vector and, using word embeddings, generates descriptive captions word-by-word.
The model is trained on the dataset, which contains:
8,000 images with five captions per image.
Text files detailing training, validation, and test splits along with captions.
VGG16 Model: Simonyan, K., & Zisserman, A. (2015). Very deep convolutional networks for large-scale image recognition. arXiv. https://arxiv.org/abs/1409.1556