Image context analysis for use in social media



Journal Title

Journal ISSN

Volume Title


Tartu Ülikool


In our digital world some of the shared images have been misunderstood and such events can negatively affect person’s life and mental health. Therefore, before a person shares an image, it would be beneficial to know if the image is appropriate for the intended use. To address this problem, this thesis aims at understanding the context of the images in social media. The context of the image is described with two methods: Image tagging and Image captioning. Afterwards, a large language model is used to understand if the image is appropriate for personal, social or business use. And in this way, the person will be more aware of images shared in social media.



image tagging, image captioning, LLM, text classification, social media