Nowadays textual data on social media has significantly enhanced. Emotions can now be graphically portrayed in text-based digital communication thanks to emojis. Emoji in digital communication improve communication and create new avenues for creativity and exchange by making the text more visually appealing. Despite considerable optimisations to the neural network model for text-based emoji entry prediction, the inexperience in predicting future emoticons from images makes the task more challenging. Emojis are a great alternative to creating a sentiment-aligned embedding because they are reliable, linguistically independent, and have their own sentiment signal. Textual descriptions of picture models have garnered more attention than symbolic description models, which are now in use. In order to predict emoji from photographs, researchers used CNN architecture for image classification and an emoji2vec embedding into the word2vec model. To predict possible emoji labels, we also ran a sentiment analysis on the text. Our method successfully conveys the relationships between the emojis. This approach has been used to optimise the search time for incoming image-based emoji predictions.

Emoji, Convolutional Neural Network, Textual descroptimizeiption, word2vec model.