Our model generates descriptive captions for the given input image related to cricket domain. The proposed approach uses CNN-RNN model. For training the model, the dataset is created by us. Our model uses DenseNet201 as CNN model. This model can be further developed to automate the live cricket commentary. An image caption generator model automatically generate descriptive captions for images.
Generating well-formed sentences requires both syntactic and semantic understanding of the language.
In the context of cricket images, the model is trained to analyze and understand the different elements of cricket-related images, such as the players, the equipment, the stadium, and the audience.
Our aim is to create a description for the images specific to the cricket domain.
The proposed model is composed of two main parts : i. Convolutional Neural Network (CNN) ii. Recurrent Neural Network (RNN).