An Efficient Image Caption Generation Using GoogLeNet

Artificial intelligence researchers have shown a lot of interest in image caption generation. The ability to train a machine to describe a picture or environment like humans do has major implications in robotic vision and commerce and in many fields. Throughout the study of artificial intelligence, this has been a difficult task. Image caption generator is a technique for determining the context of an image and generating a caption for it. The existing system extracts features using convolutional neural network (CNN). The Flickr dataset is used to train the CNN model. An long short-term memory (LSTM) model will be used to build the image caption using the extracted features. One of the CNN models used in the proposed model, GoogLeNet, is used to extract features. When compared to CNN, the GoogLeNet model's performance may result in a reduced error rate. These features extracted will be fed in to LSTM, which will generate the caption for the image.